[Nav-users] Database problem

-- mvh Morten Brekkevold UNINETT >From morten.brekkevold at uninett.no Thu Nov 9 10:53:09 2006 From: morten.brekkevold at uninett.no (Morten Brekkevold) Date: Thu Nov 9 10:53:24 2006 Subject: [Nav-users] Garbage in alertengine.pid In-Reply-To: 200611090714.kA97ENQU070237@barnetv.cc.uit.no References: 200611090714.kA97ENQU070237@barnetv.cc.uit.no Message-ID: 4552FA85.5080507@uninett.no Ingeborg Hellemo wrote, On 09-11-2006 08:14: > Freshly installed NAV 3.1.1. FreeBSD 6.1-RELEASE-p10 > > /# /usr/local/nav/bin/nav status > networkDiscovery: Up > > ... > > pping.py (pid 65207) is running... > ps: process id too large: 1162996552 > alertengine (pid 65218) is running... > > /# cat /usr/local/nav/var/run/alertengine.pid > 65218 1162996552 > > The first number in the pid-file is indeed the pid, but why the garbage at the > end? Ah. I discovered that myself while creating the bug you just pointed out ;) I got tired of AlertEngine's incessant verbosity when start/stopping and querying for status, so I changed its init-script to use the shell-script based pid-file probing. The "garbage" is actually a unix timestamp, referring to when the alertEngine daemon was started. When querying the alertengine program directly (as in pre-3.1.1), it would use this to tell you how long it had been running. Why on earth the author put it in the pid file, I don't know. I thought about removing the timestamp completely from the pid file as I changed the init script, as it bothered me somewhat - but in my Linux-centric world it worked fine without removing it. Looks like I'll have to reconsider... -- mvh Morten Brekkevold UNINETT >From morten.brekkevold at uninett.no Thu Nov 9 11:17:28 2006 From: morten.brekkevold at uninett.no (Morten Brekkevold) Date: Thu Nov 9 11:17:46 2006 Subject: [Nav-users] Garbage in alertengine.pid In-Reply-To: 4552FA85.5080507@uninett.no References: 200611090714.kA97ENQU070237@barnetv.cc.uit.no 4552FA85.5080507@uninett.no Message-ID: 45530038.7090207@uninett.no Ingeborg Hellemo wrote, On 09-11-2006 08:14: >> /# cat /usr/local/nav/var/run/alertengine.pid >> 65218 1162996552 >> >> The first number in the pid-file is indeed the pid, but why the garbage at the >> end? Morten Brekkevold wrote, On 09-11-2006 10:53: > I thought about removing the timestamp completely from the pid file as I > changed the init script, as it bothered me somewhat - but in my > Linux-centric world it worked fine without removing it. Looks like I'll > have to reconsider... Ingeborg, if you're still in the patching business, you can grab r3724 from Subversion and apply it (or the ports maintainer could do it) to fix this problem. -- mvh Morten Brekkevold UNINETT >From Ingeborg.Hellemo at cc.uit.no Thu Nov 9 12:58:06 2006 From: Ingeborg.Hellemo at cc.uit.no (Ingeborg Hellemo) Date: Thu Nov 9 12:58:23 2006 Subject: [Nav-users] missing FROM-clause in databaseHandler.php Message-ID: 200611091158.kA9Bw6bs077593@barnetv.cc.uit.no The problem of "missing FROM-clause" errors when using postgres 8.1 has been mentioned on this list earlier. The workaround is to add the line add_missing_from = on to postgresql.conf, but the solution is to write correct sql-queries. If somebody wants to fix one of these errors, it is located in /usr/local/nav/apache/webroot/alertprofiles/databaseHandler.php on line [868] pg_exec() [function.pg-exec]: Query failed: ERROR: missing FROM-clause entry for table "accountgroup" --Ingeborg -- Ingeborg ?strem Hellemo -- ingeborg@cc.uit.no (Univ. of Troms?, Norway)