[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] condor_history slow
- Date: Fri, 10 Mar 2006 13:57:28 -0600 (CST)
- From: Marco Mambelli <marco@xxxxxxxxxxxxxxxx>
- Subject: Re: [Condor-users] condor_history slow
that was the binary, I copied the wrong line the log is 600MB (600K would
have been not even big...)
condor_hi 21351 root 3r REG 8,1 591370531 9224265 /opt/condor-6.7.9/local.tier2-02/spool/history
schedd at the moment is quite free (~10 jobs running), but probably it is
the time that takes with a history file around 600MB.
I wanted just to have an idea of times experienced by other people.
On Fri, 10 Mar 2006, Steven Timm wrote:
On Fri, 10 Mar 2006, Marco Mambelli wrote:
In a VDT installation un a linux machine is it normal to have
condor_history taking more than 2 min to answer?
That is about how long mine takes with condor-6.7.13. It depends
on how busy your schedd is at the moment, how many jobs in the queue, etc.
#time condor_history 173517.0
((ClusterId == 173517) && (ProcId == 0))
ID OWNER SUBMITTED RUN_TIME ST COMPLETED CMD
173517.0 usatlas1 3/8 11:13 0+00:05:18 X ??? /home/usatlas1/
The condor history file is big but not huge:
-rwxr-xr-x 1 root root 663080 Jul 14 2005
that's the binary. Actual history file lives in $LOCAL_DIR/spool/history.
450MB for my cluster and we have been up about a year on the
same history file.
I know that quill may be a solution. condor_history is rarely used, only
for debugging. Was just curios about which response time I should expect.
Condor-users mailing list