[GE users] mysql path + arco issues
crhea at mayo.edu
Mon Feb 16 20:39:15 GMT 2009
On Fri, Feb 13, 2009 at 08:53:29AM +0100, aja wrote:
> http://gridengine.sunsource.net/howto/arco/arco_mysql.txt ?
> You can find useful information there.
> Especially, do not forget to configure MySQL server to use InnoDB as the
> default storage engine!
A question and an issue for the Arco gurus...
1. Back when we installed Arco, the install guide made little distinction
on the engine used for MySQL. From Aja's comments above, it appears that
this is a big deal. What features/engine characteristics is Arco
using where InnoDB matters over the default (MyISAM)?
(My current cluster/Arco uses MyISAM and I want to understand if
it's worth it to convert the tables to InnoDB...)
2. If dbwriter dies (for whatever reason), we've seen that the cleanup process
never catches up. In poking around, we see (for example):
SELECT hv_id FROM sge_host_values WHERE hv_time_end < '2008-02-16 14:00:00' limit 500
as a sub-select of a DELETE. Why is the SQL code structured this way?
On a small-medium cluster (100 nodes/400 CPUs) the sub-select can take
several minutes to run (our sge_host_values table is 110M rows).
Why limit each sub-select to 500? When we get into this state, the
cleanup never catches up and thrashes the DB. I go in by hand and
run the delete with a WHERE clause (no sub-select, no limit).
Cristopher J. Rhea
Mayo Clinic - Research Computing Facility
200 First St SW, Rochester, MN 55905
crhea at Mayo.EDU
To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].
More information about the gridengine-users