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

Re: [Condor-users] POOL_HISTORY_MAX_STORAGE



Hi

 

Any chance this can get looked at so that we can store stats for > 1 month?

 

It appears that the POOL_HISTORY_MAX_STORAGE applies to the whole

directory, and makes the assumption that there will be 27 viewhist* files

so therefore assumes that if all files reach max size then any one file

can’t be > 66.7 Mb in size?

 

Thanks

 

Cheers

 

Greg

 

P.S. the silence was deafening from my previous post J (below) so should I

be sending this to condor-admin instead?

 

From: condor-users-bounces@xxxxxxxxxxx [mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Greg.Hitchen@xxxxxxxx
Sent: Wednesday, 4 April 2012 4:09 PM
To: condor-users@xxxxxxxxxxx
Subject: [ExternalEmail] [Condor-users] POOL_HISTORY_MAX_STORAGE

 

We keep pool history info in our Condor setup on our Condor ViewServer machine.

This is a standalone machine that collects info from 5 separate Central Managers.

 

For historical reasons we had forced condor to represent each machine as one

“resource”, i.e. NUM_CPUS=1

 

We have recently enabled core detection and now have a total of ~ 10,000 cores

in across all 5 pools. In recently using condor_stats to produce some monthly

stats it’s become obvious that information is being lost, i.e. it appears we don’t

have info going back far enough (> 1 month).

 

Just increasing the POOL_HISTORY_MAX_STORAGE doesn’t work (currently set

at 2000000000 = 2 Gb, increased to 20000000000 = 20Gb) as we get the following

error message in CollectorLog.

 

04/04/12 13:21:05 ERROR "POOL_HISTORY_MAX_STORAGE in the condor configuration is

out of bounds for an integer (20000000000).  Please set it to an integer in the

range -2147483648 to 2147483647 (default 10000000)." at line 1693 in file /home

/condor/execute/dir_30458/userdir/src/condor_utils/condor_config.cpp

 

From what I can see our viewhistory directory is only 737Mb in size (see below).

There does seem to be some forced file rotation though at individual file

sizes of ~66.7Mb. Can anyone confirm what’s meant to happen with these

storage limits and file sizes and rotations?

 

Thanks

 

Cheers

 

Greg

 

>ll

total 736780

-rw-r--r-- 1 condor condor 31434648 Apr  4 15:49 viewhist0.0.new

-rw-r--r-- 1 condor condor 66666695 Aug  9  2011 viewhist0.0.old

-rw-r--r-- 1 condor condor 41337118 Apr  4 15:37 viewhist0.1.new

-rw-r--r-- 1 condor condor   333359 Apr  6  2006 viewhist0.1.old

-rw-r--r-- 1 condor condor 10537682 Apr  4 15:21 viewhist0.2.new

-rw-r--r-- 1 condor condor   333380 Jan 19  2006 viewhist0.2.old

-rw-r--r-- 1 condor condor 42825820 Apr  4 15:49 viewhist1.0.new

-rw-r--r-- 1 condor condor 67153008 Apr  4 09:36 viewhist1.0.old

-rw-r--r-- 1 condor condor 27489661 Apr  4 15:37 viewhist1.1.new

-rw-r--r-- 1 condor condor 66981236 Apr  3 23:24 viewhist1.1.old

-rw-r--r-- 1 condor condor 35099274 Apr  4 15:21 viewhist1.2.new

-rw-r--r-- 1 condor condor 66884552 Mar 31 17:43 viewhist1.2.old

-rw-r--r-- 1 condor condor  1208195 Apr  4 15:49 viewhist2.0.new

-rw-r--r-- 1 condor condor 66666869 Mar 27 18:51 viewhist2.0.old

-rw-r--r-- 1 condor condor  1889444 Apr  4 15:37 viewhist2.1.new

-rw-r--r-- 1 condor condor 66666889 Feb 15 05:28 viewhist2.1.old

-rw-r--r-- 1 condor condor 17508889 Apr  4 15:21 viewhist2.2.new

-rw-r--r-- 1 condor condor   333437 Mar  7  2006 viewhist2.2.old

-rw-r--r-- 1 condor condor 41038505 Apr  4 15:49 viewhist3.0.new

-rw-r--r-- 1 condor condor 66666970 Nov  4  2010 viewhist3.0.old

-rw-r--r-- 1 condor condor 27010376 Apr  4 15:37 viewhist3.1.new

-rw-r--r-- 1 condor condor   333372 Mar 15  2006 viewhist3.1.old

-rw-r--r-- 1 condor condor  6818397 Apr  4 15:21 viewhist3.2.new

-rw-r--r-- 1 condor condor   333371 Mar 13  2006 viewhist3.2.old

-rw-r--r-- 1 condor condor        0 Sep 12  2005 viewhist4.0.new

-rw-r--r-- 1 condor condor        0 Sep 12  2005 viewhist4.1.new

-rw-r--r-- 1 condor condor        0 Sep 12  2005 viewhist4.2.new