| |
| ESXi 4.1 and the disk capacity... Update! |
| As a follow-up to the article I wrote on July 22 (ESXi 4.1 does not use full disk capacity for vmfs3) an important information has to be updated! Today I installed another ESXi 4.1 server, using also a Dell PowerEdge server (but this time 2950) with a total disk capacity of 1.4TB. And surprise surprise, this time ESXi formatted the vmfs up to the full size, so the vmfs3-partition has a size of 1.4TB!  This means that the bug, described on July 22, is only occurring when the total disk capacity is bigger than 2TB!
|
| Yes, you have maintenance, but not for batteries |
| One of the main reasons to accept a maintenance contract with hardware vendors like Dell, HP, IBM, etc. is to NOT buy new/replacement equipment, in case some parts become defect. That's normal right? Well it seems that this doesn't count for batteries in Dell PowerEdge servers anymore. I received several hardware warnings which led to the conclusion that the raid controller battery (Romb Battery) failed. After opening the case with the Dell Pro Support, the technician announced that batteries aren't covered in the maintenance contract anymore.
Excuse me?! A Pro Support maintenance contract costs several hundreds of dollars a year and is usually never used, and when it comes to a failed battery part you don't want to replace it? DELL, if you want to keep your company clients satisfied, then do whatever you can so your customers keep paying a lot for your maintenance contracts but continue to replace EVERYTHING which fails in a server. Without exception.
|
| iSeries/AS400 Disk Expansion Installation |
| Today was playtime again! The newest goodie to install: A Disk Expansion Bay for IBM's iSeries.     
Once racked and plugged through SAS connection, the disks had to be
- Initialized and Formatted
- Include to existing raid group
- Added into existing ASP This was done just in the correct moment... My nagios status from this morning before adding the new disks:
Storage: OK - 20.77% free
Which is just 0.77% away from a degraded iSeries (after 80% used there is a loss in performance). New nagios status now is:
Storage: OK - 68.24% free So for some months it will be calmer... until the same situation arrives!
|
| 100 |
100! It took 821 days or 2 years, 2 months and 30 days until this post!
Originally launched as a reminder for some work-specific tasks, I realized pretty fast, that some HowTo's were well visited - so I continued writing HowTo's when I thought it might be helpful for others. Meanwhile the website has an average of 80 unique visitors per day having reached the so-far-maximum of 102 visitors a few days ago (August 4 2010).
And since the SEO optimization coding I made earlier this year (April 2010) the numbers continue to grow as you can see from the following statistics:  From post #1 until this post there were
- 14'576 visitors
- 22'771 pageviews (which makes an average of 1.56 pageviews per visit)
- visitors which came from 147 different countries Thanks to all the people having visited and will continue to visit this website!
|
| The tabulator in bash |
I've just spent interesting 20mins with my boss, trying to figure out why Bash wouldn't recognize our command to not output lines containing tabulator entries. The problem: grep -v "\t" tempfile # does not work
Still shows lines with tab-entries. At first we thought maybe Bash has a problem recognizing the tabulator, so my idea was to use sed instead of grep: sed -e '/\t/d' tempfile # works
It works with sed! So the \t is correctly recognized. Maybe a problem of grep only? After doing several tests, we finally found the solution. First we used an echo to force a tab sign within the grep command: grep -v "$(echo -ne "\t")" tempfile # works
So as soon as a sub-shell is opened to 'translate' \t into a tab-sign it works, how it is also written in the bash-manpage: Words of the form $'string' are treated specially. The word expands to string, with backslash-escaped characters replaced as specified by the ANSI C standard. Backslash escape sequences, if present, are decoded as follows:
...
\t horizontal tab
...
Which leads to the following and final solution:
grep -v $'\t' /tmp/tempfile # works |
| Flying Cat Part II |
It feels like it's a déjà-vu...
As I wrote EXACTLY one year ago (how crazy is that?!) our cat Lenny fell down from the 3rd floor. The story ended well, just so you know. Meanwhile we have a second cat called Twix and what happens today? Yes, you already know it... this time Twix fell down. Exactly one year after Lenny. One year ago it was pretty easy to catch Lenny as he was sitting shocked on the ground - but not with Twix. He was even more scared than Lenny was a year ago and when he saw me he ran away and was hiding under a balcony.
As if I wouldn't have better things to do in the mornings at 7am I somehow climbed under the balcony, had to approach him in soldier-style moves (on all 4 and pressed to the ground) through several spider nets, ant hills, and itchy plants and then I finally got him, still in shock. But he's alright.
Noting down: Don't let cats out of the window on August 19th's!
|
| HowTo Change Confixx email account password in Roundcube |
One of Roundcube's extensions is the 'password' plugin which allows the logged user to change his mail password. Several solutions exist, e.g. to change the password via a SQL query, via LDAP, chpasswd, and so on. In this case the goal is to change the password of a user managed by Confixx and where the email account's password usually only could be changed by the Confixx user.
The new HowTo 'Change Confixx email account password in Roundcube' explains step-by-step how to activate the plugin and which sql statements to use to succeed in this. Enjoy and have a good weekend.
|
| Connection type fix in new version of check_equallogic |
In the newest version of the Nagios plugin check_equallogic (20100809) there has been a change in the check of the ISCSI connection (type conn). Due to a wrong assumption all the given connections were taken as one. This works fine if only one Equallogic (one member) is used, but as soon as several members are in a group, the numbers wouldn't be correct anymore.
The new version resolves this issue and the total number of ISCSI connections of the group is now used. You can download version 20100809 here.
As always, feedback is very welcome.
|
| Veeam Backup & Replication 4.1.2 now likes Virtual Center 4.1 |
As I previously wrote Veeam's Backup & Replication application in version 4.1 (and 4.1.1) didn't really like the upgrade of VMware VCenter Server 4.0 to 4.1. From that day on, replication jobs didn't work anymore.
But now Veeam has released a new version (4.1.2) which solves that compatibility problem. In the release notes this is indicated as:
"Support for VMware vSphere 4.1. Veeam Backup & Replication now supports VMware vCenter Server 4.1 and VMware ESX(i) 4.1."
After some tests I can confirm that: Replications work again!
|
| Windows for public |
| On my daily trip into the data center I have to walk next to a commercial center in the heart of Geneva. Every day I see some commercials running up and down on a big screen running on big LED's (around 5cm diameter). But today it was different. Additionally to the endless loop of commercials I saw something I knew: A grey Windows XP task bar! 
Well what can I say about this? Not much really, but it would have been friggin funny if I saw a Bluescreen on it! |
| ESXi 4.1 does not use full disk capacity for vmfs3 |
Strange behavior yesterday a newly installed ESXi 4.1 machine: The vmfs3 partition used only around 27% of the disk capabilities. Meaning: From the logical harddisk of 2.86 TB only 744GB were used for VMFS. After some extended research and several tests, it looks like the ESXi 4.1 installer may have a bug in the partitioner, causing to add a primary partition for vmfs which is smaller than it could be. But with manual tweaks on the ESXi machine and (if necessary) in a partition tool like GParted, one can get rid of this bug by manually recreating the vmfs partitions. I wrote a How To called VMware ESXi 4.1 does not use whole disk capacities for VMFS3 which describes in details how to manually repartition and reformat the vmfs volumes on an ESXi 4.1 machine. Update 29.07.2010:
Meanwhile I've contacted VMware to get some information from the 'source' why ESXi is doing this. The official response is that this is a 'wanted limitation' of ESXi and it is unlikely that this limitation will disappear in the near future. So if you want to use ESXi 4.1 with a lot of disk capacity you either may have to switch to ESX (for manual partitioning) or you follow my HowTo which serves as a workaround here.
|
| Veeam Backup & Replication 4.1 doesn't like new VCenter servers |
Last week I upgraded a VCenter Server from version 4.0 to the newest version 4.1. Version 4.1 now needs a 64bit operating system (yes, it's a must now) and therefore I reinstalled the physical server with the 64bit version of Win 2003.
I started the VCenter Server inventory from scratch again, adding all the physical ESX and ESXi servers to it. And I wondered, if Veeam Backup & Replication 4.1 would correctly work again... the VM's names didn't change so if it works by vm object name, there shouldn't be a problem.But no. The Veeam jobs didn't run anymore and ended with an error: Analyzing object "VM-Name" Object VM-Name not found. There are no objects to backup What can be done about that is to go into the properties of each job and reselect the VM's to backup. This adds the new so called refID from the new inventory of VCenter Server into the Veeam database and should now communicate with the correct VM. Important Update 20.07.2010:
Even though some manual adjustments were made in the job properties and even in the SQL database, the jobs won't run. After having contacted Veeam support, the answer is simple: Veeam Backup & Replication in version 4.1 is not compatible with vSphere vCenter Server 4.1!
|
| When Windows 2003 setup doesn't find the (SAS-)disks... |
Was it really that long ago? Have 4 years passed so fast? That is around the last time I had to install a new Windows 2003 Server from scratch on a Dell PowerEdge server. Back then it was a Power Edge 1850, yesterday it was a 1950. Pretty much the same history: Windows setup can't find the hard-disks and therefore aborts the installation.
4 years back I simply used a diskette with the LSI Logic drivers for SAS disk and of course I had this diskette ready on my desk for future use. Meanwhile I am at another company and all my hardware-preparations (CD's and diskette's) are gone. And now the bad news: The diskette link on the Dell site has disappeared! But here's what to do now, in 2010: 1. Download the Dell USB Key F6 Driver Utility, v.1.2.4, A08 (or newer) and install it (e.g. into C:\Program Files\Dell\USBKeyPrepF6). 2. Download the Dell Driver (not Firmware!!) for your SAS Raid controller from. This is in most cases the PERC 5/i or 6/i depending on your server model. The easiest method is to go to http://support.dell.com, search your server model, select your wanted operating system (watch out between 32bit and 64bit!) and open the section SAS Raid Controller. Then click on the 'Dell - Driver' link and select the Hard Drive version of the driver. To help you find it quicker:
Windows 2003 32bit: DELL_MULTI-DEVICE_A07_R211422.exe
Windows 2003 64bit: DELL_MULTI-DEVICE_A07_R211424.exe
 3. Execute the downloaded driver in point 2. A self-extract setup will launch and select the following folder to extract the files into: C:\Program Files\Dell\USBKeyPrepF6\Files (depending where you have installed the Dell Utility in point 1). 4. Now depending on your PERC version (which you can see on your boot screen of the server) you might need to change the file :\Program Files\Dell\USBKeyPrepF6\Files\txtsetup.oem. Follow the table below for which cases the line starting with 'scsi =' has to be changed:
Storage Controller TXTSETUP.OEM needs to be changed?
Dell SAS 5/E Adapter No change needed
Dell SAS 5/i Integrated No change needed
Dell PERC 5/E Adapter Change to: scsi = PERC_32
Dell PERC 5/i Integrated Change to: scsi = PERC_32
Dell PERC 6/E Adapter No change needed
Dell PERC 6/i Integrated No change needed
LSI Logic CERC ATA 100 Change to: scsi = PERC32
LSI Logic PERC 4/DC Change to: scsi = PERC32
LSI Logic PERC 4/Di Change to: scsi = PERC32
LSI Logic PERC 4/SC Change to: scsi = PERC32
LSI Logic PERC 4e/DC Change to: scsi = PERC32
LSI Logic PERC 4e/Di Change to: scsi = PERC32
LSI Logic PERC 4e/Si Change to: scsi = PERC32
LSI Logic PERC3/DC Change to: scsi = PERC32
LSI Logic PERC3/DCL Change to: scsi = PERC32
LSI Logic PERC3/QC Change to: scsi = PERC32
LSI Logic PERC3/SC Change to: scsi = PERC32
LSI Logic LSI2032 No change needed
Dell SAS 5x/6x Controller No change needed
Adaptec U320 HostRAID Controller No change needed
Adaptec CERC SATA 2s Controller No change needed
Adaptec 39320A Controller No change needed
Adaptec 39320A Leadfree Controller No change needed 5. Now plug a USB memory stick up to 16GB into your computer and launch USBKeyPrepF6.exe. Warning: All the data on the memory stick will be deleted. The utility copies the files and formats the USB stick in a special way, which makes the stick able to be seen like a virtual floppy drive when booting on a Dell server. Don't worry, when you can't access the stick anymore by Windows, that's normal now. 6. Plug the memory stick in your Power Edge server and boot it. You have of course also inserted the Windows 2003 installation CD. So you boot from the CD and when Windows setup asks you at the very begin to hit the F6 key (for additional drivers) then hit it. After loading some drivers, setup is now asking if it should look for additional drivers. Hit the 'S'-key to search and the PERC drivers will be found on the plugged memory stick. That's it! :-) |
| Zip-Unpacking problems with Amavisd-new 2.6.4 |
Today my Anti-Spam-Virus-Gateway running with Postfix/Amavis/ClamAV ran into a strange problem. It let through a virus in a exe file, which was packed as a zip file. When I installed this gateway, I tested this behavior of course and there were no problems at all. So there are two questions to be answered: 1) Why did the packed exe file go through (it should have been banned in the first place)?
2) Why wasn't the virus detected? To answer the first question, I did some research in the logs which show that the zip file couldn't be uncompressed: Jul 7 14:59:19 linux amavis[15983]: (15983-02) (!)Decoding of p003 (Zip archive data, at least v2.0 to extract) failed, leaving it unpacked: Compress::Raw::Zlib defines neither package nor VERSION--version check failed at (eval 79) line 467.
Now it makes sense, that the mail with the exe was not banned since Amavisd couldn't decompress the zip file to see that there is an exe in it! But why does that happen? It has worked before. After some more and intensive version checkings and research I figured it must be due to a newer Amavisd-new version which I installed lately. And indeed, it is. I removed the current rpm amavisd-new-2.6.4-28.1 and installed the older version amavisd-new-2.6.3-8.1.x86_64.rpm:
rpm -e amavisd-new-2.6.4-28.1
rpm -ivh amavisd-new-2.6.3-8.1.x86_64.rpm After that I resent from an external e-mail account a mail with a zip attachment containing an exe file and see what happens in the mail log: Jul 7 16:51:58 linux amavis[32264]: (32264-01) p.path BANNED:1 recipient@example.com: "P=p004,L=1,M=multipart/mixed | P=p003,L=1/2,M=application/x-zip,T=zip,N=windirstat1_1_2_setup.zip | P=p006,L=1/2/1,T=exe,T=exe-ms,N=windirstat1_1_2_setup.exe", matching_key="(?i-xsm:.\\.(exe|msi|dll|vbs|pif|scr|bat|cmd|com|cpl)$)" That's how it should be! So the problem comes definitely from Amavisd-new 2.6.4. Or at least from the rpm package for SLES10. To answer the second question. This is a virus which was detected in the wild only today. This can unfortunately happen. For this case I'll install another virus scanner next to ClamAV with another signature base following the principle two eyes see better than one. |
| New version 20100705 of check_esxi_wbem |
I would like to announce the immediate availability of check_esxi_wbem.py in version 20100705. Aaron Rogers reprogrammed the Exit Msg/Exit Code logic on the HP server check. Before, the exit code was OK even though a critical message and exit code should have been the case (in his case a disk was defect). His correct logic has been added to version 20100702.
Since this logic was added, the plugin would only return UNKNOWN states for Dell servers which are OK. Therefore I added the same logic at the end of the Dell server check as well (20100705). This has been tested successfully on both HP and Dell servers. |
| Hardware Status tab in vSphere client disappeared |
This weekend I updated a lot of ESXi servers to the latest version (build #261974) with the Host Update Utility 4.0. This went really smooth, no problems at all. Even bigger was the surprise, when I suddenly couldn't find the 'Hardware Status' tab in vSphere Client anymore.
A quick research on Google pointed me towards the Plugin-Manager (Plug-ins -> Plug-in Manager) and there it was: The vCenter Hardware Status plugin was disabled: 
Looking at the error message, it seems that the vSphere Client couldn't correctly fetch the scriptConfig.xml file using port 8443 from the vCenter Server.
Another research pointed me to the information VMware KB 1010641: Enabling the Virtual Server Service status and hardware status plugins fails. This KB information confirmed it. One of the points in the checklist is to verify, that the computer on which vSphere Client is running can connect on port 8443 to the vCenter Server.
To solve it: Check your firewall settings and verify that port 8443 is open. If you can still not connect via a simple 'telnet vcserver 8443' don't blame the firewall yet. On the vCenter server check if the 'VMware VirtualCenter Management Webservices' service is running. It could be that this service is set to a manual startup type so you have to start it manually. Once this service is started, check the connection to port 8443 again. It should work now. |
| New version 20100628 of check_esxi_wbem |
I would like to announce the immediate availability of version 20100628 of check_esxi_wbem. This version contains the following features or fixes:
- Plugin outputs server information
As we know it from other hardware plugins (like check_hpasm or check_openmanage) the check_esxi_wbem.py plugin now outputs server information as well (server model, serial number and bios information)
- Unknown set as default exit code
The previous default exit code was 'OK'. So if something didn't work correctly, the plugin states 'OK' anyway. This has been fixed.
- Correct 'Authentication error' message
Before, just a not very informative 'CRITICAL' error was shown when the authentication to the ESX/ESXi server was not correct. This patch now shows a correct 'Authentication error' message in the Nagios output.
All these patches and implementation of the new features were created by Samir Ibradzic - all credits go to him! Thank you a lot for your contribution. |
| Exchange says 550 5.7.1 authentication required and rejects mail |
Since Exchange 2007 there is a new behaviour in Microsoft's Exchange Server for distribution groups. If an e-mail comes from an external server (e.g. from internet) the e-mail might be rejected and the sender might never receive an error message. This results in confusion on both sides: Where the hell did the mail go?!
The resolution is simple. The Exchange logs (Message Tracking) show the following error message with an EventId FAIL:550 5.7.1 RESOLVER.RST.AuthRequired; authentication required By default, distribution groups are "made for internal purposes" and not for e-mails received by internet. Open the Exchange Management Console and open Recipient Configuration -> Distribution Group. Right-click on your distribution group and select Properties.
 In the 'Mail Flow Settings' tab click on 'Message Delivery Restrictions' and click on Properties.
You will see, that 'Require that all senders are authenticated' is activated. Because emails from internet cannot be authenticated (by Active Directory obviously) they will not be accepted. As soon as you deactivate this option, e-mails from internet or other external sources will be accepted. |
| Hands on Dell PowerVault 124t (LTO5) |
LTO5 is slowly but surely rolling into enterprises and with never-ending-growing data, this is a good, adapted way to make backups. One LTO5 tape can hold up to 1.5TB of data - even 3TB compressed. That's a lot of capacity for one tape... But to be able to use the new LTO5 tapes, a LTO5 tape drive is necessary. One of them is the Dell PowerVault 124t and here are the pictures:     As you can see, there is one ethernet interface (1Gb/s) and one SAS-connection for a physical server or workstation (which sends the commands to the PV124t). The new SAS connection is necessary for LTO5 tapes, at least on the PV124t, otherwise the data transfer would be too slow. By default, the PV124t still comes with a SCSI card, but as said, if you want to use LTO5, you need to change that.
|
| check_confixxlicense does not check Confixx MAC-address |
Just a quick 'one more thing' about the Nagios plugin check_confixxlicense... Please note, that this plugin currently only checks the validity of the date. I ran into problems last weekend when I had to change a physical server. Therefore a new MAC-address of the eth0 interface was in the system and Parallels obviously stores the MAC-address together with the Confixx license on a registration-server. A try to reactive the license in the Confixx menu won't work: > *(3) Die Seriennummer ist bereits für einen anderen Server
vergeben. Bitte überprüfen Sie Ihre Daten. * So as of now, the plugin only checks the validity of the license using the timestamp of the end date. Just wanted to have it mentioned.
|
| I don't believe you! |
If someone is at a bakery and wants to buy a (overpriced) sandwich, I guess it is normal to expect a correct change in return. If some cents are missing and the customer kindly gives a hint, that the change was not correct, the employee definitely shouldn't do a conversation in the following way:
I gave you the correct change!
- You gave me back a 10 Swiss Franc bill, twice a 2 Swiss Franc coin but there's still missing 20 cents.
I already gave you that!
- Ma'am, I have your change in my hands.
I don't believe you! Maybe I should use this strategy at work as well. If someone calls with computer problems and I get a YES! on my 'Reboot-Question', I just shout 'I don't believe you!'. Although I doubt, that this strategy is very well thought of and will work... |
| The iPad has arrived - and this is serious! |
The iPad has made its way to Europe and is now also available in Switzerland. Every technician knows already in advance about the lacks of the iPad, but we have to admit, the marketing of Apple has reached a big number of people willing to buy the new flagship.
How serious this all is, shows this Apple Store in the center of Geneva, where not less than 7 security guards are standing at the exit of it. 
No, I didn't go in. But if I would have, I guess the number of security guards would at least double. Man, this is serious! Oh and by the way: The picture has been taken with my Android phone :D.
|
| New version 20100526 of check_equallogic available |
There is a new version (20100526) of check_equallogic available for download now. I'd like to thank Roland Ripoll for his hint, that the plugin didn't work correctly on a FreeBSD system because of improper usage of the snmpwalk command. It became a bad habit to me, to set the options at the end of the snmpwalk command; the options should be set before the hostname of course (as it is even written in the very first line of the snmpwalk --help output). Shame on me!
The funny thing is though, that it seems, no Linux system didn't really care about how the snmpwalk command is being used. Another thing with FreeBSD, here the commands will not be executed correctly. Anyhow, this is now corrected and works now also on FreeBSD.
|
| Swisscom ADSL started to block port 25 |
In the past years, more and more internet service providers started with the unsocial behavior, to block the standard SMTP port (25) - mostly without notifying the customers. I've always been very skeptical about this blocking and I was glad it seemed that there are working real technicians at my current ADSL provider (Swisscom) who know about the importance of using standard (and RFC!) ports. Until yesterday. While trying to send an e-mail from my Android phone (with K9 mail app) yesterday using my Wireless LAN at home, which uses the ADSL connection from Swisscom, I received the following error message from Swisscom, I have to change ports to 587. 573 Swisscom Antispam: Authentifizierte Verbindung nicht moeglich.
Bitte benutzen Sie den Port 587 anstelle von Port 25. Weitere
Informationen: www.swisscom.ch/p25.
Connexion authentifiee pas possible. Veuillez utiliser le port 587 a la
place du port 25. Ulterieurs informations: www.swisscom.ch/p25.
Collegamento autenticato non e possibile. Si prega di utilizzare la
porta 587 invece di porta 25. Altra informazione: www.swisscom.ch/p25.
Authenticated connection is not possible. Please use port 587 instead
of port 25. More information: www.swisscom.ch/p25.
Luckily I'm a technician so I know what this means, but now imagine this for the rest of the customers of Swisscom. 95% wouldn't know what this means and would start to blame anything else. Oh and by the way - this only affects the ADSL/fixed line connections yet. So far no problems to send e-mails on port 25 via 3G/EDGE/HSDPA. |
| VMware: vMotion stops at 10% and fails with Invalid fault |
Some time ago I had some problems with a new configuration of VMotion. As soon as I started a VMotion of a running virtual machine, the process stopped at 10% and returned an 'Invalid fault' error: 
I knew I've seen such an error a long time ago, when I did my first VMware HA setups in 2004 but I quite couldn't remember the exact cause of it. This summarized task list helped me to remember it. The simple cause is that the ESX hosts must resolve the DNS shortnames using the /etc/hosts file. So if you run into trouble, you have to add the server's short names into /etc/hosts like this: 127.0.0.1 localhost
::1 localhost
10.1.2.141 ESX01
10.1.2.142 esx02.localdomain.net esx02
10.1.2.143 esx03.localdomain.net esx03
10.1.2.144 esx04.localdomain.net esx04
10.1.2.145 esx05.localdomain.net esx05
10.1.2.146 esx06.localdomain.net esx06
10.1.2.147 esxi1.localdomain.net esxi1
10.1.2.148 esxi2.localdomain.net esxi2
10.1.2.149 esxi3.localdomain.net esxi3 After this, vmotion was working fine! |
| Wyse support: Frustration guaranteed |
support case , the; su|pport case; [a systems engineer needs help, which is already a rare case, opens a support case (also known as support ticket) at a Hardware or Software vendor; reports a bug; expected to receive a solution or at least answer within a short period of time respecting SLA] I already wrote a few times about some support cases. Some of them were handled great and very professional. And some of them were just sending useless nonsense as a reply... but in general if one opens a support case, an answer will follow. Not with Wyse. Since I work with Wyse ThinClients (V10L) I usually was able to solve all problems myself. 3 support cases were opened since July 2009. One of them was basically only a question, so not a real support case. The second case concerned a software problem which was causing high CPU on virtual machines having Wyse software installed. And the third one was opened two weeks ago filing a bug report or requesting investigation because the Wyse TCX Multi-display made (and is still making) financial java applications crash. Seeing the following status of the cases just makes me angry: 
The only case which was resolved was the little question I had. The response contained a direct link where to find a specific download because the website doesn't link to it.
What the hell means a cancelled case? The problem was never resolved and it seems as some time passed (July to October 2009) they just cancelled the request. Of course without sending a notification e-mail.
The newest support case is now open and being declared as second-highest priority (medium business impact) it didn't even get assigned within 9 days. Thats 2 weeks in business days! The 'Last updated' info just shows, that I have changed the priority in hope to finally get someone taking care of my case. Wyse might produce some good pieces of ThinClients - but their support is ridiculous.
|
| New (modified) Nagios plugin to check hardware on ESXi |
This is actually not a new Nagios plugin at all, it is a modification of the check_esx_wbem.py script, written in Python by David Ligeret. But yesterday I ran into the following two problems using the original plugin:
- For a host I received only an output (null)
- The hardware check didn't at all work on a HP ProLiant 460c where a Battery issue was reported on vSphere Client On the internet I found two newer versions, one which solves the first problem, another one which solves the second one. To use both advantages you either have to use two different checks with both different versions of the plugin OR you combine both modifications in one script. And that's basically what I did. But I also added a hardware type switch (either 'dell' or 'hp') to determine the type of the host server. In my tests I sometimes got an OK message when there was a hardware problem due to different responses from HP and DELL servers. So this switch is necessary - at least for certain hardware failures. All you need now is the plugin: check_esxi_wbem.py |
| What might be the best Blade system? |
Whenever I look at my web stats, I'm overwhelmed and shockingly surprised, how many people find its way to my site using the very informative keywords i hate blade servers. To use such keywords in a search engine means it must have caused a lot (hell a lot) of frustration.
Top 5 keywords from January 1st 2010 until May 4th 2010
Now the question is: Which blade vendor actually causes most trouble? Unfortunately I can't compare this in personal working experience but good old friend Google helps us a bit.
If one searches for sun blade problem Google reveals that there are approximately 512'000 results. Let's do this on four big blade server vendors and this is the result:
 
| sun | dell | hp | ibm | | blade problem | 512000 | 1060000 | 2110000 |
775000 | | blade error | 152000 |
565000 | 1080000 | 413000 |
I know this actually doesn't represent real error statistics and doesn't prove which blade system is better or more stable, but it is funny anyhow to see the enormous differences in the amount of web pages concerning blade problems. |
| HP re-releases ESXi Offline Bundle 1.3 |
Honestly, what is going on at HP?
As I wrote in the last post, HP released a version 1.3 of its ESXi Offline Bundle (on April 27) and deleted the download and the page again the following day.
Now it seems HP released the ESXi Offline Bundle 1.3 - again. This time they call it Version: 1.3 (a) (30 Apr 2010). This (a) looks a lot like Alpha to me but HP recommends users update to this version at their earliest convenience. There's nothing to see in the release notes why this is called 1.3 (a) either.
I guess I just have to try and install it again. Hopefully this time without the errors. |
| HP ESXi Offline Bundle 1.3 - not working and disappeared? |
Followers of this blog -not many but at least a few- know, that I'm not a very big fan of HP. The more surprising it was to me, that a new software release of HP made me happy yesterday. I'm talking about a new version of the HP ESXi Offline Bundle (1.3). For those who don't know what this is about, this includes several CIM informations for better and more detailed hardware monitoring with vSphere Client. I was surprised and happy, HP makes some effort to satisfy customers and offers more ESXi support. And it has just been released April 27th so it's hot news!
But enough enthusiasm... It's HP!
Today when I tried to install it, I got the following error message on vSphere CLI:
vihostupdate.pl --server 1.2.3.4 --install --bundle hp-esxi4.0uX-bundle-1.3.zip
Enter username: root
Enter password:
Vib signature is missing.
Vib signature missing? What the..? It looks like the bundle was not correctly created. I tried to go on the same download site again to find some additional info and here is what I got:
We apologize, we were unable to find the driver description document you requested.
Subscribe to drivers and support alerts to be notified when information is added.
The web page has been deleted. If you search on the HP website, the latest ESXi Offline Bundle is now (again) 1.2. This is proof to me, that this package has not been tested well, has been released and that HP got too many complaints during two days, so they removed the download again.
Once more... frustration thanks to HP. Sad!
Update 4.27pm: After hitting a refresh on the ESXi software overview, the latest available version is now officially back to 1.2. Good I made a screen-shot before! This confirms my theory, that version 1.3 was released even though it was buggy (so not well tested!). |
| Follow-up on using NTLM on Windows 7 for Proxy Authentication |
As I previously wrote in article Using NTLM on Windows 7 there's need to set a new registry key to enable NTLM on Windows 7:
This is the registry key you have to set yourself:
In HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa create a new DWORD value named LmCompatibilityLevel and set the value to 1. Reboot.
This is correct, however this only works, if the affected Windows 7 machine is member of a domain (Active Directory). For standalone machines in a workgroup the following step is also necessary:
 Run Local Security Policy -> Security Settings -> Local Policies -> Security Options.
Change the settings for 'Network security: LAN Manager authentication level' -> set it to 'Set LM & NTLM - Use NTLMv2 session if negotiated' (the default is 'Not Defined').
Once this settings is in place as well, you can use your browser and authenticate with your proxy server. |
| Follow-up on Connection Error when trying to send mail on Android |
This is a follow-up on the previous post Connection failed error when trying to send mail on Android.
One possible fix for this problem, as written in the filed bug report 1063, is to install the K-9 mail application, which according to the author should work fine.
I tested it and installed the K-9 application from the Android market - without success. Even K-9 sent a localhost helo.
But.. what I didn't try, was to install the latest version of K-9 which is still in development and not yet released in Android Market. I installed K-9 rev 2703 from the from the K9 project website and tested to send an e-mail using K-9. With the following result:
gprs03.swisscom-mobile.ch[193.247.250.3]: EHLO [10.142.41.98]
gprs03.swisscom-mobile.ch[193.247.250.3]: 250-area-1.ch
gprs03.swisscom-mobile.ch[193.247.250.3]: 250-PIPELINING
gprs03.swisscom-mobile.ch[193.247.250.3]: 250-SIZE 10240000
gprs03.swisscom-mobile.ch[193.247.250.3]: 250-VRFY
gprs03.swisscom-mobile.ch[193.247.250.3]: 250-ETRN
Success! K-9 tries to send the localhost value as HELO command, if that fails, it uses an internal IP (here 10.142.41.98).
I finally can use my Android like I need it to! |
| Connection failed error when trying to send mail on Android |
It's been bugging me for weeks. And it hasn't changed since. No, I'm not talking about volcanic eruptions on Iceland. Neither am I talking about my little cat Twix which is still not able to open the little cat-door so the cat-toilet has to stay inside... No, I'm talking about the problem of the Android Email application which can't send mails.
If you try to send an e-mail using your Android Email application (not Gmail!) you may see a yellow bar which contains the disappointing text CONNECTION FAILED. Why is that, you may ask yourself. Should I have bought an iPhone instead, may be the next question. The answer to all that is to analyze the problem.
On the remote mail server, the try to send the e-mail is clearly refused because of a helo check:
Apr 22 15:53:04 area-1 postfix/smtpd[23187]: NOQUEUE: reject: RCPT from area-1.ch[78.46.75.44]: 554 5.7.1 : Helo command rejected: you are not me; from= to= proto=ESMTP helo=
A debugged log of Postfix shows the following connection coming from Android:
gprs47.swisscom-mobile.ch[193.247.250.47]: 220 area-1.ch ESMTP Postfix (Debian/GNU)
watchdog_pat: 0x559270
gprs47.swisscom-mobile.ch[193.247.250.47]: EHLO localhost
gprs47.swisscom-mobile.ch[193.247.250.47]: 250-area-1.ch
gprs47.swisscom-mobile.ch[193.247.250.47]: 250-PIPELINING
gprs47.swisscom-mobile.ch[193.247.250.47]: 250-SIZE 10240000
gprs47.swisscom-mobile.ch[193.247.250.47]: 250-VRFY
gprs47.swisscom-mobile.ch[193.247.250.47]: 250-ETRN
lost connection after EHLO from gprs47.swisscom-mobile.ch[193.247.250.47]
The problem is, that the used HELO/EHLO command used by the Android Email application is 'localhost' which of course is a senseless SMTP helo - everybody is localhost and you don't connect to a remote host like this. Now it would be interesting to compare this with the iPhone, and here we go:
gprs31.swisscom-mobile.ch[193.247.250.31]: EHLO [10.122.235.211]
gprs31.swisscom-mobile.ch[193.247.250.31]: 250-area-1.ch
gprs31.swisscom-mobile.ch[193.247.250.31]: 250-PIPELINING
gprs31.swisscom-mobile.ch[193.247.250.31]: 250-SIZE 10240000
gprs31.swisscom-mobile.ch[193.247.250.31]: 250-VRFY
gprs31.swisscom-mobile.ch[193.247.250.31]: 250-ETRN
So here we can compare it. Android uses 'localhost' and iPhone some internal IP (10.122.235.211) which isn't perfect either but much better than localhost. Depending on the mail server configuration this might also fail but the chances are definitely greater that the internal IP gets accepted. Many mail servers however reject remote hosts trying to send e-mails and identifying themselves as localhost for a good reason; a lot of spam servers are using this method.
There is already a bug report open, issue no. 1063 handles this case. But unfortunately without resolution.
I'll follow up on that and try to see what happens, when the entries in /etc/hosts are changed... |
| New version of check_equallogic released |
Hello folks!
There's a new version (20100420) of check_equallogic available right now. It allows now to check more than 3 ethernet interfaces with the ethif check-type. For example a PS6000XV uses 4 ethernet interfaces and the original ethif check only monitored eth0, eth1 and eth2.
I'd like to thank Michael Goetze for this hint and his beta-testing of the new version.
Thanks also to everyone for all your feedback I've received by e-mail. |
| Why -while read line- does not work with stdout |
For some reason I was stuck in further developing of check_equallogic because of one little bugger: I couldn't get shell outputs (several lines) into one array. This has been frustrating since it didn't make any sense to me, why Bash wouldn't correctly read the lines into an array.
Until I understood.
My mistake:
i=1
cat /etc/resolv.conf | while read line; do myarray[$i]=$line; (( $i+1 )); done
It might seem correct - but it isn't. Thanks to this forum topic I finally understood that 'while' opens its own shell and is not able to read the values: The while statement is part of a pipeline and runs in its own shell so the array value is not available to the main script..
So the solution is done with a for loop (why didn't I think of that before, ouch!):
i=0
for line in $(cat /etc/resolv.conf); do myarray[${i}]=$line; i=$(($i + 1 ));done
And that's actually why I love working in the IT sector: You never finish to learn (= it never get's boring)! |
| New Nagios Plugin: Check Confixx License |
To all you ISP and webhoster admins out there. Ever got into the problem, that your Confixx license has ran out? I know you could check this manually every day in your Confixx or you could schedule a cronjob which renews your license every few months - but has anyone had the idea to actually monitor the license and be notified before it ends? No? Well now there is!
I present a new Nagios plugin created these days and thought of over the last couple weeks. It's called check_confixxlicense and does not more and not less than check the current validity for your Confixx license.
Documentation and Download can be found in the HowTo section on page Nagios Plugin check_confixxlicense Confixx License. |
| Google Gears a Virus (Trojan.Downloader-90750) ? |
Today our Squid proxy server denied the download of GoogleGears.exe saying it found a virus (Trojan.Downloader-90750). At the begin I thought, this must be a false-alert by ClamAV but by doing a refresh of the download site, the download worked and after another refresh the same virus was found again.
Here what the squid logs are saying:
1) Download blocked - Virus found:
1269416489.415 372 internal IP TCP_MISS/200 12879 GET
http://dl.google.com/update2/1.2.183.23/GoogleInstaller_en.application?
squid.user DIRECT/74.125.79.136 application/x-ms-application
Wed Mar 24 08:28:43 2010 [29891] LOG Redirecting URL to: http://squid.
ip/clwarn.cgi?url=http://dl.google.com/tag/s/appguid%3D
%7B283EAF47-8817-4c2b-A801-AD1FADFB7BAA%7D%26iid%3D%7BF5B2DDAE-
AC72-6B11-E4D9-E2ADF4A58E0B%7D%26lang%3Den%26browser%3D2%26usagestats
%3D1%26appname%3DGears%26needsadmin%3DTrue/gears/
GearsSetup.exe&source=internal.IP/-&user=squid.user&virus=stream:
+Trojan.Downloader-90750+FOUND
2) Download worked, not blocked, no virus:
1269420614.010 26 127.0.0.1 TCP_MISS/200 379 HEAD
http://dl.google.com/update2/1.2.183.23/GoogleInstaller_en.application?
- DIRECT/74.125.79.190 application/x-ms-application
1269420614.060 50 127.0.0.1 TCP_MISS/200 12879 GET
http://dl.google.com/update2/1.2.183.23/GoogleInstaller_en.application?
- DIRECT/74.125.79.190 application/x-ms-application
1269420614.164 182 internal.IP TCP_MISS/200 12879 GET
http://dl.google.com/update2/1.2.183.23/GoogleInstaller_en.application?
squid.user DIRECT/74.125.79.190 application/x-ms-application
Note the different remote servers of Google: 74.125.79.136
(Virus found), 74.125.79.190 (download without problem). I tried to contact Google but as everybody knows, it is difficult to get in touch with the admins over there. We'll see if there was really a virus on one of the Google servers. |
| PHP Compilation fails after upgrade to Debian lenny |
I was just about to do a compilation of the latest PHP 5.2 version (5.2.13) on the recently updated Debian server (from etch to lenny) when I saw several compilation errors. So it seems, some - imho too many - dev-packages were not updated. The following packages had to be newly installed (they were not in the system anymore):
aptitude install apache2-threaded-dev
aptitude install libbz2-dev
aptitude install libcurl4-gnutls-dev
aptitude install libjpeg62-dev
aptitude install libpng12-dev
aptitude install libfreetype6-dev
And if you receive the following error message, as I did, you have to install the following package.
configure: error: utf8_mime2text() has new signature
aptitude install libc-client2007b-dev
Cheers |
| Dear LiteON support |
Dear LiteON,
I hope that someday, someone of your direction or your marketing dept. or even someone of the support has got five minutes of spare time, googling for 'LiteON support' and will finally land on this page. Because this is intended to be an open letter.
When you create an optical DVD drive iHAS424 with typical SATA connections and you bundle it with a SATA Slim Data+Power adapter, have you ever tried to connect this male adapter to the female connections on your drive? Because I you have ever tried, you would have seen that they are NOT compatible at all.
My desperate attempt to inform you about this mistake by writing to your support was followed by unnecessary and completely -sorry- stupid answers written in poor English saying basically 'this is not an IDE drive'.
Thank you! Actually this is WHY I bought this drive, because it is NOT an IDE drive.
 But as I can quote your first answer on my demand 'Could you send the photo of the SATA-adapter to us so that we will check whether the drive bundled with correct accessories?' I already knew where this whole support request was going - Nirvana, Neverland, Nowhere, Wonderland (because I really wonder how your support works). You are the technical support and you do not even know what kind of adapter you bundled.
And even for my friendly answer sending you your wanted picture (since my description obviously didn't create a picture in the mind of the technician) I got a ridiculous answer saying 'the bundled SATA Cable and SATA power cord is correct'.
If you are able, with brute force, to connect a DVI monitor to a VGA graphic card, then you are also able to connect the bundled adapter to your iHAS424 drive and your answer must be correct.
Thank you again very much. |
| $* is no longer supported - Confixx Counterscript error |
... and we're not completely done yet with the Confixx Counterscript problem. Well as described in the previous post, it runs again but the following warning is still shown:
$* is no longer supported at /usr/local/majordomo/majordomo.pl line 47.
$* is no longer supported at /usr/local/majordomo/majordomo.pl line 47.
$* is no longer supported at /usr/local/majordomo/majordomo.pl line 47.
This error is also due to the perl update to 5.10. As this forum post reveils majordomo is using old (deprecated) tags in perl. But more important, you will find this link which shows you exactly which lines to edit in majordomo.pl.
Now the Counterscript is running smoothly again. |
| Confixx Counterscript fails after Debian Upgrade (Lenny) |
I recently updated two servers from Debian etch to lenny which in most cases ran smoothly (detailed info will follow). After testing all the services (FTP, HTTP, and so on) I supposed that the upgrade is finally done and everything works as wanted again. Mistake! The Confixx Counterscript didn't run anymore and showed the following error message:
Can't locate Confixx/Filter.pm in @INC (@INC contains: /etc/perl /usr/local/lib/perl/5.10.0 /usr/local/share/perl/5.10.0 /usr/lib/perl5
/usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl .) at /root/confixx/confixx_counterscript.pl line 2.
To find the solution how to solve these error messages and to let the Confixx Counterscript run again, check out the HowTo Confixx Counterscript fails after upgrading Debian to Lenny. |
| The HP Blade nightmare continues |
Déjà vu.
This is the expression used in English and German for an event, which already happened exactly like this (or which is felt as having already experienced). In French it's a simple sentence saying 'already seen'. Both cases describe perfectly my HP Blade experience.
As I already described in an old blog post (dated Nov 30th 2008) I've had tons of problems with HP Blade servers. The nightmare seems to continue.
New company. Newer blade system. Still HP. Still problems.
For several months now several Blade servers show a degraded status, caused by memory problems. Once I described the problem to HP's technician, he admits that the Enclosure might have a problem and that everything has to be upgraded. I mean everything, the whole damn thing: Internal switches, ILO/BIOS/Firmwares of each Blade server and the Enclosure firmware itself. Meaning: Several downtimes.
Downtime? Wasn't that the word used by the HP vendor visiting us last week in combination with the word 'no'? I remember slightly having heard the words 'no downtime' considering the Blade server design/redundancy and its enclosure.
One good thing I can say about that at last: This is nothing new to me and I'm not at all surprised. |
| HTTP Forward using path variable with JavaScript |
HTTP Forwards are being used a lot. And it's even very easy to use the actual path if your domain or URL has changed for example.
So let's say you are using a free hoster and you want to forward your webpages to another domain:
http://username.mywebhost.bla/files/page1.html -> has to be pointed to http://www.mydomain.info/files/page1.html
Note that only the domain name changes.
The problem in this situation was, that mod_rewrite and PHP is not activated on the free hoster server (username.mywebhost.bla). So we have to find another solution.
Instead of changing each page of the website and set a meta forwarding to the new URL (this could take a while depending on the size of your website!) there is still the ability to use a JavaScript code.
You can let JavaScript make the output of the meta forwarding, altogether with the actual path. This looks like this:
<script language='javascript'>
var path = location.pathname;
document.write("<meta http-equiv='Refresh' content='0; url=http://www.mydomain.info"+ path +"'>");
</script>
Place this into your HTML HEAD section and it works. JavaScript MUST be enabled in the browser of course but still, you get a nice and working dynamic forward even when you can't use PHP or .htacces Rewrite Rules. |
| Changes in Sun's Customer Support since Oracle took over |
So here we go again. As I already let out my frustration about a previous SUN support case, my bad experience with SUN seems to continue.
Well everybody in the IT-world is aware that Sun has been overtaken, flagged, purchased, slaughtered (and killed?) by Oracle. But I didn't imagine in my dreams that SUN would just stop to execute customer orders or to satisfy clients.
In fact, I just opened a hardware support case with Sun, describing that a sensor of a StorageTek has detected a dying battery which has to be replaced. No problem I thought.
-> Contacted Sun's Customer Support by e-mail.
-> Answer from Sun after 40mins, saying the maintenance contract ended.
-> Another e-mail to apologize we forgot to check that and if I could have an offer for the spare part (or that my request gets forwarded to the concerned dept)
-> Final answer from Sun: "bedingt durch die Uebernahme durch die Firma Oracle bieten wir diesen Service seit dem 01. Februar nicht mehr an. "
Translated to English: "Due to the take-over of SUN by Oracle we do not offer this service anymore since February 1st."
Now isn't that great?! Thanks a lot! >:/ |
| VMware: Schedule automatic reboot or shutdown of physical ESX(i) server |
Today, the first time for a long while, I am again in the need to make a scheduled shutdown of a physical ESXi server. Strangely it seems that this function disappeared out of the vSphere Client. I'm pretty sure, it worked with one of the first VirtualCenter versions.
But after a bit of research I found some hints about the vSphere CLI SDK. After a bit of testing I figured out how it works and it works just great.
Now I'm able to schedule a cronjob, launching the reboot or shutdown command fully automated. To be able to shutdown or reboot, you have to enter maintenance mode. The command looks like this:
/usr/lib/vmware-vcli/apps/host/hostops.pl --server VCENTER_SERVER --target_host NAME_OF_ESX_HOST --operation enter_maintenance --username VCENTER_USER --password VCENTER_PASSWORD
Note, that for the --target_host you mustn't use the IP-address of the ESX/i server but the name used in the VirtualCenter configuration (as you see it in your vSphere Client).
Now we make the same for the shutdown command:
/usr/lib/vmware-vcli/apps/host/hostops.pl --server VCENTER_SERVER --target_host NAME_OF_ESX_HOST --operation shutdown --username VCENTER_USER --password VCENTER_PASSWORD
Commandline baby! |
| Argh - already a bugfix for delapacheuserfiles.php |
Well I've redone some of the tests of the PHP script delapacheuserfiles.php today - with the result of finding some bugs in my own script. D'oh!
One of the most important bug is the following:
When you launch the script and the root-folder (the lowest folder which is owned by the apache user) is not chmodded 777, the sub-files and sub-folders within will be deleted BUT this folder itself will stay as it is. Even the permissions will stay the same, so it's not even possible to delete it via FTP.
I've now changed the handling for the lowest apache-user-owned folder. In some cases it won't be deleted, but the permissions will be changed so the FTP user can still delete it.
New updated version can be found in IT Howto's / PHP Script: Delete files and folders created by Apache user. |
| How to delete files created by Apache user? |
There is one problem on shared hosting servers with Apache and PHP running on it: If a PHP script creates files and folders the owner of those created files is the user under which Apache is running, that's mostly www-data or wwwrun. User's accessing their webaccount by ftp and wanting to delete such files receive a permission denied error - because (of course) they're not allowed to delete those files.
Here an example of an ls -l output:
-rw-r--r-- 1 www-data www-data 0 2010-01-16 18:30 file1
-rw-r--r-- 1 www-data www-data 0 2010-01-16 18:30 file2
-rw-r--r-- 1 www-data www-data 0 2010-01-16 18:30 file3
drwxr-xr-x 2 www-data www-data 4.0K 2010-01-16 18:30 testfolder1
drwxr-xr-x 2 www-data www-data 4.0K 2010-01-16 18:30 testfolder2
drwxr-xr-x 2 www-data www-data 4.0K 2010-01-16 18:30 testfolder3
So if you're someone who was seeking for a solution when you can't delete the files with your ftp user, this is it!
Find more information in IT Howto's / PHP Script: Delete files and folders created by Apache user. |
| Gmail now supports sending from external e-mail addresses |
As the technical skilled Gmail (Google Mail) users know, one of the few disadvantages of Gmail is the use of external e-mail addresses.
Example: If I wanted to send an e-mail to someone with an e-mail address say mailtest@claudiokuenzler.com via my Gmail account, the recipient would see in the mail header that the mail came from my gmail address. Even worse: If someone uses Outlook, in the e-mail it would be written as following:
From: myaddress@gmail.com on behalf of Claudio Kuenzler [mailtest@claudiokuenzler.com]
That just doesn't look professional, right?
Today I discovered that Gmail has changed its settings page (Settings->Accounts and Import). You are now able to add (or edit) an e-mail address where you may set your own smtp server. This way, outgoing e-mails sent on your Gmail account pass via your real e-mail server, where your domain is hosted. Yippee!! |
| Installing Zend Optimizer on PHP 5.2.11 and Apache worker |
I've spent a lot of hours these last two days trying to install Zend Optimizer on a Linux server (x64-64) with Apache 2.2 using the worker MPM mode. The actual installed PHP version on this server is 5.2.11.
After downloading the newest Zend Optimizer (version 3.3.9) from www.zend.com, you may follow the installation doc but it won't work. So I found hints that Zend Optimizer isn't compatible with a) the Apache2 worker and b) PHP version higher than 5.2.9. Those hints only affect Zend Optimizer versions below 3.3 though, so it should have worked with 3.3.9.
However, after not receiving any answers in the Zend support forum I continued my research. I've tried it with a way older version (3.2.6) but here I got the problems with the worker module. With a little trick I was able to download the version 3.3.0 from the Zend website (hidden download) and what a surprise: This version works!
Summary: If you want to install and use Zend Optimizer on your Apache 2.2 webserver (with MPM worker) and PHP 5.2.11 you currently have to use Zend Optimizer version 3.3.0 (not the newest one!) |
| Dell's official answer and bugfix of Equallogic plugin |
So I'm back from my vacation and I have received a lot of e-mails from users of the Equallogic plugin for Nagios. As I previously wrote, there is a problem with checking an Equallogic which has several members - the check responses were not programmed to fetch several answers.
I've as well asked Dell if that behavior is normal or if it is a bug and they assured me on a phone conversation that this is the "architecture of Equallogic" to give answers for all members.
Luckily I don't even have to tear out even more hair off my head - I already got someone who fixed the plugin to support multiple members.
Here I'd like to thank Mathias Sundman for his work adapting the plugin to support multiple devices and Scott Sawin who was the first one mentioning this problem and did a lot of tests for me.
You will find the updated plugin in the IT Howto section. |
| check_equallogic with more than one member |
In the last few days I've gotten some e-mails from other sysadmins who are using the check_equallogic Nagios plugin. They informed me, that some checks don't work when there is more than one member in a group. After some checks, it looks like an Equallogic device responds for ALL other members as well, even if you check the eth0-address directly.
I contacted the Dell support to get some further information about that, if this is supposed to happen. Seems strange to me (as well as to the other admins) but let's wait for the official answer from Dell. |
| Good support, bad support |
Hardware defects are never really funny - but you anyway have to handle them and get the machine back to "normal" as soon as possible. The declaration of "asap" is very different, depending on which company you're dealing with.
In the last couple days I've had two different support cases, one with SUN and one with DELL - and hell! What a difference!
I've opened the case with SUN on November 6th, gave information about the serial number and type of the server, described the problem (hardware fault) and delivered the Linux kernel logs describing the crash. I had to wait 10 (!!!!) days to get an answer saying I have to upgrade the BIOS and other hardware firmwares. I did as said the following day with the same result: Server crashes. And since then I'm again waiting for a response.
The second case concerns DELL. My Nagios plugin (check_equallogic, soon generally available) detected a Disk failure. I confirmed this on the Group Manager and really, a disk was showing status failed. I've opened a support case this morning at 9:17am and guess when the disk was replaced and case 100% solved? A few minutes ago at 12:38pm.
To make it a bit more fair, I have to say, that the DELL support is a 4h contract and the SUN support is a Next Business Day contract, but still - I guess you can imagine how much time it should take to solve a problem, even with a Next Business Day case. |
| Using NTLM on Windows 7 |
Since not even one month Windows 7 was officially released for the public and is available in stores.
Today was the first time, I actually had to work with Windows 7 and ran right into the first problem: NTLM Authentication against a Squid Proxy Server. Each time I started IE (8), a Pop Up would ask me for my login credentials. Even when I entered them, no website was shown, instead the Po Pup came back.
I checked the option in IE to allow credentials to be passed further (Tools->Internet Options->Advanced->Security/Enable Integrated Windows Authentication) and it was checked, therefore activated.
What could it be?
Then I found the following forum post: "The problem is the autentificación NTLM of Windows7. It is necessary to create the following key in the registry to solve it".
This is the registry key you have to set yourself:
In HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Lsa create a new DWORD value named LmCompatibilityLevel and set the value to 1. Reboot.
After the reboot I started IE and without even asking me anything I could start browsing. Hurray! |
| Nagios Plugin for Dell Equallogic |
I've recently searched for a Nagios plugin which checks the status of a Dell Equallogic - without any success. Probably the first time, I haven't found a plugin. Neither on NagiosExchange (or now MonitoringExchange) nor somewhere else via Google.
Solution? Write the plugin myself! I will keep you posted when I'm finished. |
| Wintertime |
As the time changes, conditions change as well
mankind to adapt, choosing heaven or hell
Early darkness, fulfilled with everlasting nights
will take over these days, turn on the lights
Coldness will hold in, will each of us do harm
Every soul living blank, will be missing the sun
And as desperate thoughts seek sun with cries
We all know darker times in front of us it lies |
| Postfix rejects incoming e-mails with 450 Server configuration error |
I've just wondered about this error message in my Postfix logs and couldn't at the begin figure out, why I see these error messages. It happened for EVERY incoming e-mail, therefore they were temporary rejected.
I googled but couldn't find anything relating until I realized that postgrey (Greylisting function for Postfix) was not started. So if you are using Postgrey on your mailserver and you see "450 Server configuration problem" in your maillogs, check that Postgrey is actually running ;-) |
| Fighting Spam with SpamAssassin |
There are lots of Anti-Spam-Solutions out there and the past few months I've excessively digged into this topic and have set up a big anti-spam system.
SpamAssassin is one of the most used solutions against spam. Of course, SA itself isn't enough at all but it helps you out in case some spams broke through your first wall (e.g. RBL's).
Unfortunately, I haven't found many updates rulesets for SA the last months and SARE seems to be down, too. So over the past weeks I wrote my own ruleset and added several regex into it. Maybe someone out there will find this handy and wants to reuse it. Here's a brief description and the download to it.
Have fun. Fight spam. |
| Atelier2.ch is 10 years online! |
The website Atelier2.ch was the first commercial website I've ever done, at the funny age of 13 or 14. With it, I've received my first official payment and I was sooo happy with it.
Looking back at it now, it was a LOT of work, especially to do all the scanning of pictures (Digicams only just started). It was all done with static websites and with frames (!), my god, what a layout hehe.
Time passes by, I get older and got lots of more experience doing websites. That is now, 10 years later, the big opportunity to relaunch the website, and it's done!
 
The old and the new website layouts, 1999 and 2009
As you see in the picture, it looks very different now. The whole new website is dynamical with database access in the background and a fully implemented (and self-programmed) cms system. A big difference to the old one ;-).
Oh yeah, before I forget: Congratulations for being more than 10 years online now!!! Not too many websites can say that. |
| Flying cat |
I knew this would happen someday but I was always hoping it never would.
Since I live in an apartment on the 3rd floor and my cat is always running around on the edge of the balcony, I knew it is just a matter of time until he slips and will fall down. So happened yesterday night, when he was outside of the window. The ledge is rather big so I'm still surprised that he fell down, maybe he got scared of something and then fell. So he must have been "flying down" for around 3 seconds asking himself 'Hey.. Whats going on? What am I doing here? Why do I feel so light?' BUMM! And he lands (luckily) on the grass part behind the building.
At this time we realized he fell down and got a light to search for him. Indeed, he was on the ground, looking up to the window. Of course I ran down and outside immediately, searching for him, worried if he was hurt. He seemed OK but how scared he was.. But once he was in my arms he surprisingly (and annoyingly) started to get excited and he seemed to like being outside (and maybe to fly?). The whole night he was being crazy, over-excited.
I think it's an experience he made, which has changed him. And I hope he learned to be a bit more careful... and I really hope he doesn't think he can fly now! |
| A black VMware Weekend (Upgrade ESX 3.5 to ESX 4.0) |
What a black weekend that was! I know it is meanwhile Thursday but it still is worth a later blog entry.
I've planned a maintenance to upgrade 4 HP Blade servers (480c) from ESX 3.5 U2 to ESX 4.0 build 171294. In the upgrade test done last Friday everything went well so it was a GO for the production servers on the weekend.
The first two servers didn't cause any problems at all. It just took a LONG time to upgrade all the VM's to Virtual Machine Version 7 afterwards.
Let me give you a hint here: Once the ESX is up again running on ESX 4.0: Start the VM's and gather information about the current network settings. Then upgrade VMware-Tools, reboot VM. After that upgrade the virtual hardware and boot the VM again. Depending on the VM it will install new hardware once logged in and will reboot (Windows VM's did that). When the VM is finally up again, VERIFIY the network settings. It is possible that they were overridden or set to DHCP..
OK, we continue...
With the next server the problems started. After a BIOS upgrade the server couldn't boot anymore. There was obviously a problem with the motherboard and the new BIOS version detected it. HP support was called and they finally delivered the new motherboard at 1am.
The next day it was time to upgrade the last server. Everything went supposedly smooth, installation and boot, everything. Once the server booted ESX 4.0 and the server was available in VirtualCenter (now called VSphere Client), I saw two VM's which were "inaccessible". These VM's were stored on a SAN datastore which was not mounted as VMFS3 anymore. I'm still asking myself why! So I added the same LUN to another ESX server which could mount the VMFS datastore, setting a new signature. When I browsed through the datastore I saw the two VM's. Yes!
But then the shock: I did the same thing on the ESX server where it was still written "VM's inaccessible". I manually mounted the VMFS datastore as in the other ESX and it didn't get mounted! It got formatted as VMFS3! My VM's ... gone... aahhh!!!! :-/
After one hour being on the phone with VMware, the only thing they could say was "sorry" and that they didn't know why this happened. I guess it had something to do that the upgrade of ESX somehow changed the LUN identification number but the installation is not supposed to do that. That's theory though.
At least now everything is OK again since, fortunately, the VM's were not THAT business critical.
What do we learn of that? You must create a backup of your VM's before upgrading to ESX 4.0!!! Definitely do that or you might experience the same black weekend as I did! |
| Ghostbusters return! |
I recently had the chance to be at the famous Montreux Jazz Festival and followed the concert of Ray Parker Jr. Great funky tones, some Rock, great music.
At the end of the concert, of course, he played his most famous song "Ghostbusters". This groovy song moved my ass - I still like this song after so many years.
And that's when I thought: I should redo my research about a new movie! It was 2 years ago when I first did a research with a sad end: Dan Aykroyd already wanted to film a third Ghostbusters movie some years ago but ONLY with the original crew. Bill Murray always refused to play in a sequel, so the idea died.
And now, in July 2009, after that concert in Montreux, I decided to check it again and: WOW! Ghostbusters III is in production! With the same crew! The state is currently "scripting", so they're writing the script of the film. Planned movie release: 2012! Whohooo!
Who're you gonna call? |
| Eureka is back! |
Finally!
After months of wait for the return of my favourite tv series, it's finally done: Eureka is back! Hurray!!
No clue why the SciFi channel, now called Syfy, stopped the production after the first 8 episodes of season 3 - the season of the new episodes aired on SciFi are now called "Season 3.5".
If you have no clue what Eureka is, check out the official website on syfy.com. |
| Testing a new Mailserver |
I'm currently setting up a new anti-spam/anti-virus server. Therefore I have the chance to use a completely new domain to make a test...
I wonder how many spam mails I will get within the next hours and days when I now write the following e-mail address to a public blog:
paul@openair-kinos.ch
Now come here, all you spam-bots outside! :-)) |
| New mobile phone - a difficult decision |
Since the breakdown of my old SonyEricsson V800 in 2007, I have been using a Qtek 9000, a huge stone called Smartphone with a Windows Mobile 2003 Edition on it. It has been in my pocket for the last 6 years and it has been running since then. But time goes on and newer/better/faster phones are available.
For the last two months I have been thinking about doing what everybody does and buy an iPhone. Or do something new and choose an Android phone. My decision was leaning towards the HTC Magic which is using the Android platform but last week, Apple announced the new iPhone 3Gs (Speed) with video, faster OS and other small improvements.
iPhone 3Gs Pro and Contras
++ Easy to use
++ Many satisfied other customers (proven in business)
++ Multi-Touch
++ Huge internal disk (16GB for the one I am interested)
++ A huge list of applications to install from AppStore
-- Only one application open (lying in front of the others)
-- Dependent on Apple's will
-- U must use iTunes
-- No SSH client (you have to "jailbreak" the iPhone)
-- Almost everybody has it
...
HTC Magic Pro and Contras
++ A new phone
++ Open-Source OS
++ AppStore which is getting more and more apps
++ Almost everything configurable
-- Small disk (512MB only, possibility to insert memory cards)
-- No 3.5mm audio plug!!! aahh!
-- No video
-- Dependent on Google (at least in some functions)
Today I have gotten the news that Samsung is releasing its own Android phone: The Samsung i7500 Galaxy. And this is the real iPhone challenger! Check this out:
Samsung Galaxy Pros and Contras
++ 8GB internal memory and possibility to insert memory cards
++ Video
++ 3.5mm standard audio plug
So everything what I thought "what a shit" done by HTC Magic will be improved with the Samsung Galaxy.
My choice is made! ^^ |
| Ready for the new challenge! |
| I've recently decided to move on with my current job situation. In the past months I was unfortunately almost not able to work with Linux or OpenSource products - something what my "business heart" is beating for and I think is very important in the IT industry. To not lose my knowledge I was working hard for in the past years I decided to move on and quit my job (that was in march). Tomorrow, 1st of may 2009, I will start my new position as Linux Engineer in an internationally well recognized trading company. There will be way more responsibilities lying on my shoulders but I am ready for it! Wish me luck and success. |
| Funny support call! |
This week I'm on call service, that means, in case of questions or problem I must answer the phone calls starting at 6pm and ending 8am. Usually it is quiet. But this morning at 7.45am I received a call. Here's the call log:
> Hello, this is Claudio from the IT.
- Yes, hello. I have a problem with my computer. I did not shut it down yesterday because someone had to create a subfolder with my 10'000 mails. Now it is not working anymore!
> OK, what is not working anymore? What do you see on your screen?
- Everything is black and there's a yellow thing moving around.
> *thinking: this must be the screensaver!* I think your screensaver is active. Did you already press a key on your keyboard?
- No.
> Could you please press a button on the keyboard?
- Ah! Now it looks normal again! Thank you!
This is not a joke, this is really what happened this morning! It's 2009 and there's still people out there who don't know screensavers. That's my personal lesson today =). |
| Windows 7 - the second try and a love story (?) |
So here's now an update to the history of the Windows 7 Beta installation which crashed (see the last blog entry).
The failed installation was actually caused by a defect DVD which I've received from a Microsoft employee. It was an original Beta DVD, not a copy, so I thought "that must work".
I've now downloaded the Iso-file and burnt a completely new DVD and installed Windows 7... with the following results:
- Installation was VERY easy and fast. After 30mins my system was complete and I could log myself in.
- The boot of the computer takes 35seconds until the logon screen appears. After the logon it took another 12 seconds until the system was ready and was not loading apps anymore. It is fast but of course the machine is empty.
- The installation of Firefox took 5 seconds - compared to a new installed Windows XP that is rapidly faster!
- The mix between shortcut icon and open program is a thing to adapt but it is a step into the good direction, it is better for the overview to not have an endless list of open programs in the task list.
- On my main pc (Windows XP) I shared the WLAN-Connection. Windows 7 made a search in the network and explored, that my main pc is now a "Residential Gateway Device". I was surprised, that this would be discovered automatically.
- The automatic change of desktop wallpapers is a gadget but nice :).
- The Control Panel has a good overview and is loaded really quickly (compared to Vista's really long listing).
So now I have to explore it a bit more. So far I can say that: I like it. It's a nice one this time! |
| Testing Windows 7 - how? |
Tonight I've told myself: Let's test Windows 7! I've recently received the Windows 7 Beta from a Microsoft seller and thought, oh well, let's give it a try.
But how the heck am I supposed to test it when it won't install?!!! After 50% of expanding installation files, I receive an error:
Install Windows
Windows cannot install required files. The file may be corrupt or missing. Make sure all files required for an installation are available, and restart the installation.
Error code: 0x80070570
Oh great! I've of course redone the installation with the same result. And it is an official Microsoft Beta DVD, not just a downloaded version... Nice start! |
| I've refound my favourite childhood movies! |
When I was young (around 6-8 years) I have seen some movies which stayed in my memories. Without ever knowing what movies they were. Today I finally started my research - with success! Thanks to Internet and the good combination of keywords I've finally found the names of these movies!
The Computer wore Tennis shoes (1969)
One of my favourite movies when I was a kid. Pretty difficult to find this one!
The only thing I remembered of this movie was, that a boy got an electro shock when touching a computer and somehow all the data information of this computer got in his brain. He then won several knowledge quizzes. The movie was from Walt Disney.
I've found the next three movies through the list of Disney feature films on Wikipedia.
The Cat from outer Space (1978)
This was pretty easy to find since there are not too many movies with a cat from outer space ;-)
Flight of the Navigator (1986)
I've had nightmares of this movie! I still remember this one scene when the boy is in the ocean and there is this huge fish trying to bite him..
The Watcher in the Woods (1980)
And now to the scariest movie Disney probably ever made! When I saw this one, I could not sleep for one week and I was horribly scared of other dimensions. Although it is in my eyes nothing for kids, it's a Disney movie and it stayed in my thoughts until now. So it must have been impressive!
There is one movie missing though and I never found the title of this movie. So if anyone could help out, I'd be glad! What I remember is that at the begin of the movie, there are robots attacking the earth. There are two different spacecrafts to evacuate and a couple is supposed to take two different spacecrafts. After the start one of these blow up. After this scene I think I had to rush to bed because my parents came home - but I still want to finish that movie in my lifetime! :) |
| How to use wired LAN and WLAN without problems |
The network installation in my apartment is not the easiest one to describe... Due to a really bad planned power and phone line architecture in this building of the 70's, I was forced to set up two different networks: A wired one for the printer and a NAS device and a wireless LAN for accessing the Internet. The problem: The wired LAN has a higher priority to connect to foreign networks so whenever I plugged the network cable, I got disconnected from Internet by WLAN.
A solution, how to use both networks at the same time without being disconnected can now be found in this new How To: Use both wired LAN and WLAN connection and use WLAN for Internet Access. |
| VMware problem |
Today I ran into a VMware problem. When using shared virtual disks between two (or more) VM's and you want to increase the size of a shared disk, the VM's will not boot anymore. You get an error message which starts like this:
Vmware ESX server cannot open the virtual disk "vmfs/volumes/.../vm1/vm1_2.vmdk" for clustering
I made a workaround to solve that problem. You will find that in the Howto: Error on boot of VM: ESX server cannot open the virtual disk for clustering. |
| The calmness before the violations (Anti WEF) |
Yesterday was the official start of the 2009's World Economic Forum (WEF) in Davos. And today, as I was waiting for the bus, I saw the following poster:
 
This poster shows without doubt what the demonstrators want to do: They want to crack the "glass of the rich", demolish it. As they announce, they want to gather in "rue de Mont-Blanc", next to all the very expensive shops and businesses. The intention seems clear: They want to break things.
In my opinion, this is once more the willingness of chaotics to vandalize something - the WEF is just a good excuse for a so called "demonstration". The spoiled Swiss youth doesn't know what to do - most of them don't even know what the WEF is doing. Surely, the WEF deserves skepticism but those people are just once more showing their disability to talk. |
| Danger in the Rush Hour! |
Yes indeed, the rush hour in the morning can be a dangerous thing! At least from what I saw this morning in Geneva's main station at 8.10am. What happened?
So the IR (Inter Regional) train arrives and several hundred people are taking the (way too!) tiny exit off the platforms. The people have to wait in front of the escalators because there is only one escalator going down - the other one is going up. When it is finally my turn I stand on the right hand side of it and taking the ride down. 2m before the end I suddenly see a woman who's got her shoe tie stuck in the escalator (as we see it usually with kids). Some men try to help her but the escalator is still running, so a bunch of people are crashing into the standing people in front. At this moment I just get trough and am able to watch the scene. I'm looking for a red "stop" or "emergency" button on the escalator, but don't see anything. A woman see's a button next to the escalator and pushes it. You'd never guess what happened! It stopped? No! The escalator doubled or almost tripled its speed! So a whole lot more people are even faster crashing into the crowd at the end. Fortunately in these seconds, a man could release the stuck tie of the woman and the crowd is falling forwards into the hall.
Have you ever seen an escalator without stop button? No? Then go to Geneva! |
| My part of Obama's Inauguration: A chocolate bar! |
Yes we can!
The USA finally got a president for whom actually people voted and not a bunch of republican judges. I've followed Barack Obama's race to the White House since the begin. It started with a thought "Hey, this guy is good!" which went on to hope "He could change America and the world!" which turned to fascination for this man, whose speeches are incredible, smart and full of emotions.
It is now two days ago, that Obama got sworn in as 44th president of the United States. The minutes the inauguration took place, I opened a big letter from my mother - surprise, there's a chocolate bar of Obama inside! At least that way I was a bit connected personally with the inauguration. Sweet!
|
| PHP 5.2.8 is out! hmmm... |
Yesterday I saw that PHP 5.2.8 was released - after last week's PHP 5.2.7 which was buggy if the magic_quotes were enabled. Oh well, that can happen. So now 5.2.8 is out and let's enjoy that! (enjoy = install, test it).
I have compiled PHP for I don't remember how many times so it's like sleepwalking to me. But yesterday it was different. What happened?
So I started the configure, followed by the make and of course the make install. At the end of the make install I got the following output on my shell session:
Wrote PEAR system config file at: /usr/etc/pear.conf
You may want to add: /usr/lib/php to your php.ini include_path
Installing PDO headers: /usr/include/php/ext/pdo/
Installing ZIP headers: /usr/include/php/ext/zip/lib
hmmm
hmmm
hmmm...? What does this hmmm mean? It's obviously a small bug according to the bug #46810 on bugs.php.net. It does not harm your PHP installation but still.. where does this "hmmm" come from? The comment of PHP Developer Jani explains it to us:
It's the braindead hack someone added for BC reasons. Pierre, it's your stuff that breaks here. Remove the crap from Makefile.frag and use the proper configure macro instead: PHP_INSTALL_HEADERS.
So that's it. In the next version, this "crap" will be removed. I love this comment! =) |
| Wrong latin1 Umlaut entries with phpMyAdmin 3.1.0 |
Last week I've received a mail from a customer complaining, that his MySQL entries are not correctly read out on his webpage (Umlaut problems, e.g. ä ü é). First I thought there is a problem with his website charset but nope, all is set to ISO-8859-1 (his website AND of course the MySQL DB/Table/Field).
He told me he does the entries not via a web form but direct via phpMyAdmin. I recently installed the latest version 3.1.0 so my guess was now, that there is a problem with this version.
After some testing with phpMyAdmin version 2.5.7-pl1, 2.11.9.3 and 3.1.0 I can now summarize this:
- 2.5.7-pl1: This old and outdated version shows the phpMyAdmin website with a standard HTML output of iso-8859-1. The Umlaut entries you do manually direct via phpMyAdmin are correct and can be read out without problems.
- 2.11.9.3: This is the newest phpMyAdmin 2 version. It shows the phpMyAdmin website with a standard HTML output of utf-8. The Umlaut entries you do manually direct via phpMyAdmin are correct and can be read out without problems (although it is done with utf-8!).
- 3.1.0: This is the newest phpMyAdmin 3 version. It shows the phpMyAdmin website with a standard HTML output of utf-8. The Umlaut entries you do manually direct via phpMyAdmin are NOT correct, the website and phpMyAdmin show some weird signes (ä instead of ä).
I've opened a thread in the German support forum of phpMyAdmin <- click. Until there is a resolution I suggest you use version 2.11.9.3 for manual inserts via phpMyAdmin on a latin1 DB. |
| Why I hate Blade servers |
In the past week I've been fighting 4 days with a HP blade server and there is no end of the story yet.
I've never liked the whole Blade idea, don't ask me why, I don't know it. But last week I got the confirmation to not like Blade systems. What happened?
First there was a server crash. A Blade server had a malfunctions and shut down. The power LED turned to red. So I call the support and instead of coming to us and repairing the server they want us to make a test with another enclosure. The next day the tech guy came and confirmed that there is a problem with the affected blade server. He took the server with him to replace the motherboard. The next day another tech guy comes by and we slide the blade into the enclosure - nothing happens. Oh! Now there is a problem with the enclosure! I got told we should reboot the whole enclosure - That means to shut down all the other blade servers as well.
Before we do that I decided to try to power up the blade server in another enclosure. The blade server powers automatically up in the other enclosure after like 15 seconds. Well, he almost powers up. The power and nic LED's are green, the disk LED stays inactive. Obviously, so I got told, there is a problem with the motherboard (again!) and a processor.
After 4 days of work we have now still 2 problems: The blade server is still down PLUS one enclosure is indicating a "Fault". Call me old-fashioned but I prefer the original use of a server with one purpose. You can still build a cluster between two different servers. If server #1 crashes, you still have server #2 in that case. But if you have a problem with the enclosure in a blade system, you risk to power off all your blade's in it. |
| Pictures of USA-trip 2008 online |
The snow has already melted but the cold stayed. What happens if you are outside in the cold wind waiting for a bus or a train? That's right: You're thinking of vacation.
So it was about time to upload the pictures of our vacations. It was kinda hard to only upload a few pictures - with two cameras we took around 800 pictures. Don't worry, you won't have to look through 800 pix, they're about 65.
So if you ever wanted to see some pictures made in Baltimore, Washington D.C., Philadelphia or New York - here's your chance.
You will find the gallery under Travelling->Baltimore/Washington D.C./Philadelphia/New York USA. |
| Snow! |
As of right now, while I am looking outside of the window, snow is still falling and by now roofs, lands and even streets are white. It looks like it's gonna snow the whole night so good luck for the traffic tomorrow :-/
It's the first time it really snows here this year and I wonder what kind of winter we will have this time. I suppose, this year it's gonna be freezing cold with a lot of snow. As long as my Mario Kart Wii works I don't care =).
Update monday morning: Something must have happened in the night. There was almost no snow anymore this morning, it was rather raining than snowing. But I saw that there's a lot of snow in the east of Switzerland where I lived before. So strange... |
| Changes in news/blog display |
If you read this, you already see the changes I have made yesterday of course. It's nothing big, the changes are simply more user friendly.
One thing I changed is the title of each blog entry. It appears now a bit bigger so it should be easier now to see where a new blog entry starts.
The second thing is a categorization of the blog entries. You are now able to select a category in the yellow highlighted row. The category you're interested in (e.g. Internet) will show all the blog entries with the topic Internet. |
| Google Mail (GMail) now with different themes (layouts) |
What a shock yesterday when I opened my Gmail account and everything was in blue! First I blamed my cat for probably having walked on my keyboard or attacked my mouse (computer mouse!) as he does that sometimes. But let's solve that.. my cat must have clicked something hehe.
I opened the settings and there was nothing to change the colors or the Gmail layout - and I didn't select anything in the labs to change the colors of my Inbox automatically.
 So I did a log out and after I logged myself in again there it was: A new feature has been enabled to use themes (different layouts) for your Gmail!
The one who has been set automatically (a dark blue one) sucks but there are really some nice ones!
Nice little gift from Google before the Christmas time, thanks! |
| VI Client can't create snapshots anymore |
Last week I ran into problems with a virtual machine to create snapshots, the buttons were simply disabled and the whole function to create snapshots seemed to be disappeared - only for this particular VM though.
After comparing the config files (vmx) of two different vm's, I figured out, it must be something with the VirtualHardware version number.
To solve this problem, just follow this howto.
Enjoy your weekend! |
| Asus EAH 2400 Pro shows videos in green - solved |
If you are someone out there who could only see green videos after installing an ASUS EAH 2400 Pro then you're helped with that now!
Under this HowTo I describe what you have to do to finally watch normal videos without a green layer again! It took me a while to figure that out... |
| The chance to prove your wiseness or stupidity |
Tuesday 4th of November 2008. Election Day.
Or, in other words: America's chance to prove that the American people learned from what happened 8 years ago and won't make the same mistake again.
The chance to prove "we learn from mistakes" or worse... to prove the stupidity of its citizens with voting for another George W. Bush (as a senator McCain voted 92% similar to Bush).
It's hard to admit but I put all my faith and hope into some million voters now. I still love the USA as a country but if its people vote again for the wrong guy.. gosh - I will think about it four times before I go visit this country next time.
Point. End. Go vote now! |
| Welcome to Switzerland, future! |
Finally future transportation systems (meanwhile old again) have made it to Switzerland. Yesterday was the public opening of the first subway in this country, called "M2" in Lausanne (in the French part of Switzerland).
Let's make a little history. The first subway (or métro) was opened 1863 in London and it still runs. The "tube" is 145 old! But let's be happy and celebrate - as yesterday some did with drinking champagne at 5.20am. Ewwah! But I have to admit, this transportation system is nice. It's way faster than the bus and you don't have to wait in the cold anymore. Plus: It looks nice (now! let's review that in 4 years again) and it gives you the feeling to finally be a part of the future.
Enough of words. Lets ride it!! |
| Cleanup Confixx after a server transfer |
Here's a little note to all of you out there who might be in a project transferring the contents of a webserver (Websites, EMails, and so on) to another webserver. The administration software is Confixx.
After the transfer we experienced the following problems:
- The customer vhosts had the wrong IP's. You can solve this problem when you update the IP addresses in the MySQL DB confixx in the tables "kunden" and "reseller".
- PHP Safe Mode is on. I'm currently still trying to get to know where this setting is coming from. As a temporary fix I set a httpd_special entry to every customer "PHP Safe Mode off".
- Some problems with the email settings of customers. But this is mainly clientside (except from 2 mail accounts which just didn't exist on the server - who know's why). |
| Microsoft Data Access Components not compatible with XP SP3!! |
Hola@everybody!
This week I've found a weird behavior of a software. The goal was to import data from an excel (xls) file to a local installed database. When trying to import the excel file an error occurred with a missing ODBC driver.
So I tried to manually install the MDAC package to give the affected machine the needed ODBC drivers. It was then when I realized that the MDAC 2.8 software is not compatible with Windows XP Service Pack 3!!
So before you install SP3, be sure you're not going to use database software which needs an internal ODBC driver.
I sent a mail to the Microsoft guys but of course didn't hear anything. Let's hope they fix that. |
| What to do when your VMware VI Client is in German |
This week I updated my ESX servers to version 3.5 Update 2 (build 110268). There's obviously a strange bug occurring when you want to connect your VI Client to the just updated ESX server: The whole language of VI Client changed into German!
My Windows is German, yes, and the keyboard layout is set to German (Swiss) but it's still a bug when you can't change the language of VI Client anymore ;-).
How to get rid of that bug will be explained in the HowTo section. |
| Back from vacation and exciting last week |
What an awesome time we had on vacation! We were so lucky with the weather (except one day with tropical storm Hanna), the US-Dollar was pretty low when we exchanged the Swiss Francs and we were right in time for the Sales in the stores!
In case you wanna go to the USA too, here are some little rules, what you should do and what you shouldn't do:
Do's in the USA
Eat cheeseburgers, take a one-hour-boat-tour (Manhattan), go shopping in Outlet-Malls, order Papa John's pizza, use the water-taxi's in Baltimore, drink Fanta Orange (way better than our Swiss Fanta!), order a big beer in Hard Rock Cafe (you get the glass to take home).
Definitely Dont's in the USA
Using public transportation systems (it takes forever, the schedule is inaccurate and sometimes there isn't even a bus!), using the USA Rail Pass offer from Amtrak (you can only ride 4 times!), be in a motel where somebody opens the ceiling above you and looks into your bathroom, have a cab driver who can't read and doesn't know where you wanna go.
Additionally for today's post: It's my last week at my current employer! I've had such a good time here and I learnt so many things. Now it is time to move on... To get into the international business. What a weird feeling, to know I'm sitting the last week in my own office. Bye bye my own office - I won't have that at my new employer. I'm ready for this new challenge! Let's rock! Hehe :-)) |
| Computer = Style |
I just received the new DELL newsletter a few minutes ago and something cought my eye: The new Desktop system called Hybrid. Computer meets Style!
Respect, Dell! That's what I call a great innovation! |
| Firefox 3 crash when trying to play a Flash video |
A real quick productive but not really long blog entry this time. Today I had to mess around with Firefox 3 (don't understand me wrong, I love that browser) and some Flash videos which made the Fox crash all the time -> whenever the video/flash plugin was loaded.
If you have the same problem with certain flash animations or websites, you may want to check out the problem solving in the HowTo section.
Keep the fox alive! ;) |
| Hot times, summer in the city |
The heat those days remind me of one song, which I like lots by the way:
Hot town, summer in the city
Back of my neck gettin' dirty-'n'-gritty
Been down, isn't it a pity
Doesn't seem to be a shadow in the city
All around people lookin' half dead
Walking on the sidewalk, hotter than a match-head
Cool - ah no, hot! - song!
But to be clear: I love this weather! I will complain in winter when it is cold again, right now I will just enjoy the sun and its warm weather!
During the last few months I was getting into the administration of Bluecoat Proxy Appliance. Because I myself was stuck with it sometimes, I wrote a HowTo for it. This HowTo explains how you can define file extension objects. Find it in the IT HOWTO'S section. |
| Actual project: Webserver Transfer |
Since my favourite two teams (Switzerland and Sweden) are now out of the Euro 2008 tournament, I can focus on my actual project. It is about transferring the content of two web servers to two different web servers. Sounds easy, but:
- Hundreds of domains have to be transferred to another name server. These nameservers will be installed on the same new servers. Of course those domains belong to customers and they have to change the nameservers of their domains. Time intensive thing and I am sure, not all of the customers know how to change the nameservers.
- 80GB of data has to be transferred through the internet to the new servers which are located in another datacenter.
- Operating system upgrade. The old servers were running on Debian 3.1 sarge, the new ones are running on Debian 4.0 etch.
- Recompile PHP and I will run into problems this time because the new servers are installed with the basic system only. So will have to install lots of packages until PHP gets to work.
- Installation and migration of Confixx will be done from a third party company since I don't really know Confixx in details.
- After all installations and data transfers, all domains in the nameservers have to be adjusted to the new ip's.
Let's begin! You will find more information when the project is finished in the "projects" page. |
| After the lost game... |
My good-mooded last entry BEFORE the game tells me, how many hopes I put into the EURO 2008 opening game Switzerland vs Czech Republic. Unfortunately, as most of you might have heard, the Czech's won 1-0.
Pictures of the way to the stadium, in the stadium and after the game can be found under travelling/Euro opening game.
For the next match tomorrow against Turkey there must be a victory, it's gonna be the last chance anyway.
Oh yeah: Congratulations to Netherlands for the unbelievable high score against Italy! |
| One day left and the game starts |
OK my friends: Tomorrow will be the day I have been looking forward to the last four years: The EURO 2008 Austria/Switzerland will start. And better than I ever imagined: I will be in the stadium! Huwahhhyeeahh!
Of course I will take pictures, of course I will show them here, of course I make you all jealous with that now ;-P.
Which nation is the best, who's gonna be Europe's best football nation? My hopes tell me Switzerland of course but that would be a huge surprise. I honestly think this year Portugal could do it. And it makes sense:
If Portugal wins Group A, they will most likely play against Germany in the quarter-final. If they win that quarter, they will most likely play against Croatia and after that in the final against France. Thats my presumption but life and the future cannot be seen in advance so we'll just see and enjoy! |
| UEFA Euro, I'm coming!! |
What a great day! At 8.15am it was already 25°C and at 10.12am my dad sends me a text message with the following content: "Wizard Merlin conjured us tickets!". Hurray!! I'm going officially to the Euro 2008! After already giving up the hope of getting a ticket into one of the matches in the stadiums, 9 days before the surprise is perfect!
And it's not just one game: It will be the opening game on saturday 7th of june in Basel: Switzerland vs. Czech Republic!
Euro, I'm coming! |
| Welcome to claudiokuenzler.com ! |
Welcome welcome to a complete redesign (and reprogramming) of my old online portfolio which was accessible on www.napsty.de.
Compared to the old website, this site includes more technical howto's in the world of IT. Whenever I will have time I will add more of those tipps. It's not only (hopefully!) useful for some people on this world who run into a problem but also a "brain reminder" for me. Yeah - we all get older! ;-)
I wrote enough now for the first entry of this new site. I hope you enjoy the site! :-) |
|
|
![[Valid RSS]](http://validator.w3.org/feed/images/valid-rss-rogers.png "Validate my RSS feed")
9997 Days
until Death of Computers
|
