Cleaning Up

The netstats backend is failing to copy the files it generates over to kahuna. This only happens when it runs as a cron job. I added some code with some error reporting to debug the problem. As mentioned preciously, my restore script for the alternate phone number ran into the same problem that we're trying to fix: inconsistent ordering of phone numbers. I wrote a script called altflip.pl that flips the regular and alternate phone numbers if my script to fix the alternate phone number guesses wrong. I could have run my script to fix the alternate phone number, but I decided doing this on a Friday afternoon was a bad idea. I'll leave it until Monday.

LDAP and Zope

In the morning I finished the program to restore the phone number fields if the alt phone change doesn't work. But after testing I concluded this was impossible, because of the same ambiguity of ordering we are trying to fix. So tomorrow I will write a "flip" script in case my logic goes screwey. Cheryl Schmidt had a problem with LDAP and I convinced her the actual behavior was the desired behavior. In the afternoon I restored a fix for DTMLTemplate that went missing when we moved to the new Zope web farm.

Alternate Phone

I've started working again on the problem of updateing the alternate phone numbers in ldap. I've reviewed my code and I decided that I need a way to back out changes, so I'm not stuck if people complain. So I started to write the scripts to do this.

Life After Netstats

I tested Harry's home page list modification and reported that it was ready for production. I wrote a script to close and open the database connections in Zope. But I couldn't get it to work, I kept running into authorization errors. Greg got his script worjing first, so I stopped working on it.

Netstats Done

I installed the netstats frontend on kahuna, did the second remedy test with Greg (which failed because of licensing issues) and did half a test of Harry's home page fix. Tomorrow I have to update the switch script, do the testing I put off, and write a database connector toggle script.

Frontend Finally Works

I resolved the problem with the cgi script. It was a permissions problem. Then I set about making the script more presentable by consolidating protocols in the plots to the most important. The final step will be to install the script on Kahuna.

Netstats Once More

I continued work with netstats. There was some problem with the backend dat files which caused the plots to look screwy. This was easily resolved. I fixed up the static "wrapper" template so it looks decent, though I'm afraid the plots might bust the layout (too large). I started testing as a cgi script. The initial form displays properly, but when it tries to compute the plots, it crashes with a 500 System error.

I also tested the consolidated project wiki for Leigh.

Netstats Works (Sort Of)

I fixed the problems with the netstats front end and it works from the command line. There are funny extra lines in the plots, though, and I'll have to figure out where they come from. After that I'll have to test it running as a cgi script and fine tune the look of the plots and web page.

There was also a permissions problem on the development system that caused some of my overnight scripts to crash. The problem was caused by the refresh from production. We've put a procedure in place so it won't happen again.

Tomorrow I have to remember to test Leigh's consolidated project wiki.

Netstats Frontend

I finished the modifications to the frontend and backend needed to run gnuplot and started testing again. There are a few problems, but nothing major. The biggest problem seems to be with the arguments to the png command and I'll probably have to run gnuplot interactively to straighten them out.

Remedy Test

The big news of the day was the lunchtime test of Remedy. This actually took longer than one hour with setup and postmortem. Before that started I modified the pingldap script along the lines of Dan's suggestions. I'll have to check back with him to get the final specs for the output. After the big test I settled on using gnuplot to generate the output plots for netstats and came up with a gnuplot script to produce what I hope are acceptable plots. I'll have to revise the perl scripts and test them tomorrow.