Today's Work

I finished recataloging the directories on gojira, though I couldn't recatalog a few because of errors. Then I checked some web pages on gojira. Some worked, some others threw errors.

Today's Work

Today I reviewed the outstanding calls and noted those I could work on, tried (unsuccessfully) to get the Palmbook to work, and started recataloging and testing the Data.fs that was moved to gojira.

Today's Work

I installed Harry's LDAP to Palm product on Ubik and started testing. I ran into problems, so I moved to something else. I looked for what was causing the problem with search, found it and fixed it, and wrote up a test plan for the fix. Then I retested Harry's STLAPManager upgrade and passed it for going to production.

Today's Work

This morning I solved the printable page shows machine url problem. I put the fix in CVS and pushed it to test. Then I tested Leigh's fix for calendar item permission checking. And we held a lessons learned meeting for the Feline brownout problem. And you'll notice that the web development news is working again.

Today's Work

This morning I watched/heped Greg work an enhancement to our call system. Then I wrote the test plan for the internalUser problem. And after that I fixed the textarea sizing problem in STLDAPManager. The I attended the hour and a half (yawn!) web coordination meeting. I got the LDAP to Palm script from Mark, which I plan to convert to a Zope external procedure.

Today's Work

In the morning Greg demonstrated how to use the Remedy developer's interface and we set up an account for me with Remedy. Then I finished installing Zope in CVS and wrote up what I did. In the afternoon I rewrote internalUser.py in Zope to minimize the number of calls to gethostbyaddr, which hopefully will fix the Zope slowdown problem.

Today's Work

I've been tagged to be Greg's backup on Remedy support and I spent most of the morning looking over his shoulder as he worked some typical problems. Then I finished getting findgaps.pl to work and ran it. I put the zope source under CVS on gojira. After that I did some research on how to solve the Zope slowdown problem and proposed killing the process and examining the traceback. Edwin came up with a likely cause to the problem, stalled DNS lookups, and we plan to increas the number of threads tomorrow to reduce the problem. Finally, Sandy asked me about a new wiki for reporting and I did some preliminary investigation.

Today's Work

I was out Monday and Tuesday because of the snow. This morning I tried unsuccessfully to import the new Zope setup on godzilla into CVS. I was unsuccessful because of some size issues with /tmp. I then turned my attention to the Zope slowdown problem. I had some conversation with Leigh and then wrote findgaps.pl to identify all the gaps in the Zope log file. It is currently running and I hope to have the complete list by tomorrow.

Today's Work

I did some stuff in the morning, but I can't remember what it was right now. In the afternon I ran replayweb against a ten minute time interval of the log file where feline had halted. I didn't notice any problems in performance. I also started adding the zope code on godzilla to cvs.

Today's Work

First, I took a peek at the Python code in ZODB to see if I could find anything related to our slowdown problem. I couldn't find anything. The John Gfeller alerted me to a problem with the LDAP user interface. I had some problem understanding his concerns, but I eventually understood that the number of rows in the form element were not sufficient. The script I wrote yesterday to replay the accesses in a web log (replayweb) cause me some problems because of authentication issues. I eventually figured out that I wasn't passing the correct string for the host. (I hadn't stripped the http://).