rblog

Blog performance improvement

I started to get preeettyyy annoyed by the bad performance of my blog. Did some research and decided just give the two lines a try in the .htaccess file


AddOutputFilterByType DEFLATE text/html text/plain text/xml
SetOutputFilter DEFLATE 


And, damn! It seems to work immediately and provide a major performance improvement. Why have I not done this before?

@Update:

Also added

<IfModule mod_expires.c>
       ExpiresActive On
       ExpiresByType application/javascript "access plus 1 week"
       ExpiresByType text/css        "access plus 1 month"
       ExpiresByType image/gif       "access plus 1 month"
       ExpiresByType image/png       "access plus 1 month"
       ExpiresByType image/jpeg      "access plus 1 month"

       # Now we have Expires: headers, ETags are unecessary and only slow things down:
       FileETag None
</IfModule>

Just now I still have some performance issues, but I suspect that to be caused by the database server struggling, not the web server.

Remove lines from file

While reviewing a SystemOut.log-file created by WebSphere I needed to remove all lines containing “SystemOut”. Luckily I have installed Cygwin, so Perl to the rescue. Found this Q&A at StackOverflow.

Ended up with

perl -ni.bak -e 'print unless m!^.*SystemOut.*!' SystemOut.log

Original file > 80.000 lines, resultfile 195 lines…a bit easier to read…

TortoiseSVN: Access is denied

Running on Windows7 I had problems with TortoiseSVN, I was unable to rename, move and such. Solution for this is to run Windows Explorer as administrator.

Initially I could also have checked out my working folder in the home directory of my own user, that should also solve the problem without the need to “Run As Administrator”

First day at IBM Hursley U.K

At IBM Hursley attending a workshop and having one-to-one meetings with WebSphere gurus together with Dino. Here just outside IBM Hursley.

At the Innovation Center, innovative clock…

Leaving IBM for the day. This picture is taken just at the roundabout outside of the area, we are really out in the backcountry

A bit further down, sheeps. Compared to IBM Hursley the main office of IBM in Norway is downtown.

Close to the end of the day, testing the Stonehenge Old Smokey, good one!

Cannot commit changes to SVN – Collabnet SVN Server

Just had an issue when trying to commit changes to SVN. Now using Collabnet Subversion Edge 3.2.2-3395.103. Have configured authentication using LDAP, that is Active Directory. Error message when trying to commit changes

svn: E175002: Commit failed (details follow):
svn: E175002: Server sent unexpected return value (500 Internal Server Error) in response to POST request for '/svn/reponame/!svn/me'

Found this very strange since it was working without problems yesterday. Tried to restart both services several times, no help. Nothing found in the logs either.

Then trying to log on to the Edge admin console with my AD-user…not possible. Using the superuser, logged in, changed LDAP from using port 389 to 3268 as the console provides a hint to do: Hint: try 3268 for Active Directory servers . Restarted Apache and voila, up an running again.

My conclusion: AD …arrghh!

FileNet – Error during installation of Workplace XT FP1 (v 1.1.5.1)

After having successfully installed Workplace XT v 1.1.5.0 applying FP1 of Workplace XT, version 1.1.5.1 fails. The error message showing in the Install Shield is plain and simple, that is meaningless: null. In the log file I found “JVM_HOME = null”.

Reported this as a PMR which helped me solve the problem. Simply just uninstalled v 1.1.5, then installed 1.1.5.1 since it is a cumulative fix. Info from IBM:

The Workplace XT 1.1.5.1-WPXT-FP001 fix pack installs a new, complete Workplace XT 1.1.5.1 system, or upgrades an existing Workplace XT system. Workplace XT fix packs are cumulative; each fix pack contains the base release and content from all previously released fix packs.

So seen, it would be possible just to scrap the existing installation including the registry path and install Workplace XT 1.1.5.1 directly.

FileNet – Workplace fails during startup

Had an issue today after having installed the application engine and then the latest fixpack. WebSphere reported that the application started just fine, but when accessing http://my_server/Workplace it failed with a NullPointerException. Part of the stacktrace

[07.02.13 18:10:45:986 CET] 0000002e SystemOut     O Loading logging config: null

This was in test, in the dev environment where it is working SystemOut.log reported

[2/7/13 18:39:24:274 CET] 00000043 SystemOut     O Loading logging config: D:\\AEShare\\FileNet\\Config\\AE\log4j.properties

After a lot of cursing it turned out that the problem was caused by Windows and access to the share where the log4j-file was located. As a wild guess I granted access to “Everyone” and voila…it worked…Windows shares…simply hate it. WebSphere running as local system account, did not cause any problems in the dev environment, but test…don’t know why yet.

Solving FNRCE0010E: E_BAD_VALUE

While preparing for the upgrade of FileNet from 4.0 to 5.1 we decided to upgrade WebSphere in the existing development environment from 7.0 to 8.0. We therefore uninstalled WAS, and left all other FileNet installation as is. While reapplying all FileNet configuration to the new WAS installation we got a problem with the Content Engine. The CE System Health page reports problems and both the p8_server_error.log and p8_server_trace.log was constantly filling up with enormous amounts of data. The errors being repeated was (showing only parts of it)


FNRCE0010E - ERROR method name: masterDecryptDecode principal name: xxxUSERNAMExxxx Global Transaction: false User Transaction.....
FNRCE0010E: E_BAD_VALUE: The value given for a property or list element lies outside the permitted range or value set...

It turned out to be caused by a change of admin uid. Initally the environment was installed using uid X, but at a given time it was decided to change it to uid Y. Since the CE ear-file was configured using X and we tried to use uid Y this was causing the error. ECM Place to the rescue, see P8 CE 4.5.1 on W2K3 x64 broke?.

List out and later reapply new user using the following commands


java -jar <CE>\lib\bootstraps.jar -e <old ear file> -l
java -jar <CE>\lib\bootstraps.jar -e <old ear file> -j <new ear file>

Hjem