[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Assertion error in Quill

Thanks for the info,

I've recreated the database, giving quillwriter access to it, and also changed the .pgpass file to contain the ip address instead of the hostname. I'm still getting the following errors:

QUILLLOG (multiple entries like this)
10/8 10:19:40 ******************************************************
10/8 10:19:40 ** condor_quill (CONDOR_QUILL) STARTING UP
10/8 10:19:40 ** /opt/condor/sbin/condor_quill
10/8 10:19:40 ** $CondorVersion: 6.9.4 Aug 30 2007 $
10/8 10:19:40 ** $CondorPlatform: I386-LINUX_RHEL3 $
10/8 10:19:40 ** PID = 25657
10/8 10:19:40 ** Log last touched 10/8 10:15:15
10/8 10:19:40 ******************************************************
10/8 10:19:40 Using config source: /opt/condor/etc/condor_config
10/8 10:19:40 Using local config sources:
10/8 10:19:40    /opt/condor/etc/condor_config.local
10/8 10:19:40 DaemonCore: Command Socket at <xxx.xxx.xxx.xxx:27037>
10/8 10:19:40 main_init() called
10/8 10:19:40 configuring tt options from config file
10/8 10:19:40 Using Polling Period = 10
10/8 10:19:40 Using logs 10/8 10:19:40 /export/home/condor/log/sql.log 10/8 10:19:40 10/8 10:19:40 ERROR "Assertion ERROR on (jobQueueDBUser)" at line 137 in file dbms_utils.C

MASTERLOG (multiple entries like this)
10/8 10:19:40 The QUILL (pid 25657) exited with status 4
10/8 10:19:40 restarting /opt/condor/sbin/condor_quill in 521 seconds

As before, condor_status and condor_q SEEM to work:

-- Quill: quill@xxxxxxxxxxxx : <xxx.xxx.xxx.xxx:5432> : quillcm
But after submitting a job, running condor_q shows an empty queue.

Any ideas?

U. of Arkansas

OK--I stand corrected... I did check before I sent the e-mail but perhaps
the old version was cached in my browser.

May I make the following coments--

on step #6 it is easy to use the wrong postgres syntax
to create the database and *think* it is owned by quillwriter
but in fact it was not.  It would help if the correct syntax
was just put there in the docs.

$ createdb --username=quillwriter quillcm -W

createlang plpgsql quillcm

I believe that the two psql loads of the *.sql files must also be
done as quillwriter user--at least that was the way it worked
for me after two previous failures and dropping the database.

Also note that there are still some bugs with how the condor_history
daemon interacts with quill--I have reported these in a condor-support
ticket.  In particular if you try to do condor_history -name quill@hostname
from some other host than hostname, it won't work.

Steve Timm