http://www-int.stsci.edu/~bsimon/weblog.cgi The work log of Bernie Simon en http://www-int.stsci.edu/~bsimon/weblog.cgi/post/311080766800.html <p> I finished coding the update to fsdiff and started debugging it. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/301080680400.html <p> I finished the changes to Mhonarc and emailed Glenn. I started work on incorporating fsdump into fsdiff. I wrote up the recovery procedure for the Zope database and put it on the <a href="http://www.stsci.edu/wiki/center/internalNews/FixProcedure">wiki</a>. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/291080594000.html <p> I spent most of the day working on the Tiger Team. It looks like we have most of the pieces for fixing the POSKey errors in place. I also did some work on Mhonarc. All problems are fixed except that the search engine doesn't generate the proper variables to populate the sidebar links. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/261080334800.html <p> I fixed the remaining problems with Mhonarc, with one exception. Much of the code in rsc.mrc will have to be moved into rsc_template.mrc, as we now have mailing list specific navigation.</p> <p>On the POSKey problem, we now have demonstrated that recreating the catalog fixes some of the POSKey errors (the famous six). We have demonstrated that fixing these errors makes it easier to repair the remaining errors. We don't have code to fix the Wiki problem other than dropping the catalog and recreating it by hand. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/251080248400.html <p> I spent the day working on Mhonarc. I redid the html template, modified the mhonarc "glue" script and rewrote rethemer. Everything works except that the list of directories on the index page is empty. Will fix tomorrow. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/241080162000.html <p> I finished the script to compare yesterday's poskey errors to today's and set it up to run as a cron job. I went back to working on Mhonarc revisions. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/231080075600.html <p> I ran a script that searches the Apache logs for specific urls. I gave a presentation on Mhonarc. And I continued to test a script to compare yesterday's poskey errors to today's. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/221079989200.html <p> I tried to come up with a method to get the object id from the oid. The folks from Zope got back to us and told us we couldn't do that. So I dropped that idea. I went on to writing a script that compares yesterday's list of Poskey errors to today's. The script is written, but not yet tested. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/191079730000.html <p> The fsrefs.py script identifies objects with dangling references. Unfortunately it produces a list of oid's, which isn't helpful to us. I spent the day looking for a way to convert oid's to object identifiers. I learned some things, but the bottom line is that I am not yet successful. </p> http://www-int.stsci.edu/~bsimon/weblog.cgi/post/181079643600.html <p> I modified the Poskey tiger team wiki page so it includes a statement of the problem and a status page. I went looking for info on the Poskey error and didn't find much new. I examined the source of fsrecover.py to see how it works. Sean found a Zope approved tool for poskey errors, fsrefs.py. I copied it to gojira, ran it and saved the output. It found may errors we were not aware of. I modifed the Mhonarc search script so that the help message is in the template instead of the script and modified the help to make it more informative. </p>