From: Rick James
Date: October 8 2012 9:35pm
Subject: RE: mysql 5.5.24/27 massive slow down since upgrade to solaris 10
 u9 with 147440-19
List-Archive: http://lists.mysql.com/mysql/228349
Message-Id: <2E7DD7ADE53B044C8C8BCD9C5829E1EB148CF9255E@SP2-EX07VS01.ds.corp.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Are you I/O bound?

XFS

What "elevator" strategy?

Compare settings.
Especially innodb_flush_log_at_trx_commit .

> -----Original Message-----
> From: Gael Martinez [mailto:gael.martinez@stripped]
> Sent: Sunday, October 07, 2012 5:46 PM
> To: mysql
> Subject: mysql 5.5.24/27 massive slow down since upgrade to solaris 10
> u9 with 147440-19
>=20
> Hello
>=20
> For nearly 3 years, we have had a master - slaves configuration working
> like a charm, the slaves would be able to sync daily ... since a recent
> upgrade from Solaris10u6 to solaris 10 u9 with the latest CPU bundle
> (kernel patch 147440-19) on one of the two slaves, the slave upgrade
> cannot keep up anymore... it is now at over 2 days behind in
> replication and the delay increases... the volume of data did not
> increase and the second slave seems in good shape (still on u6). No
> configuration change was done to the mysql configuration itself... Is
> there a well know issue/parameter to configure with the more recent
> versions of solaris ? The storage is on an Hitachi  san storage with
> 4Gig HBAs and the volume of data to replicate is about 6-8 1G binary
> log files a day... Another change is that the new machine global zone
> is using zfs for the boot disks, but the database zone itself including
> the data directory is sitting on ufs filesystems...
> Regards
>=20
> --
> Ga=EBl Martinez