Hello,
Do you have trim enabled? Maybe those stalls could happen when the
disks are getting trimmed.
Just a random thought.
Manuel
2013/6/28, nixofortune :
> Hi guys,
> Did you have any experience running MyLSQ or in my case MariaDB 5.5.31
> on FreeBSD on top of zfs?
> We are using Samsung Pro 8
Prefer xfs on RHEL.
Certain stalls are inherent in older InnoDBs, but MariaDB 5.5 should have the
Percona fixes that greatly smoothed out that problem.
What kind of drives? A RAID controller with caching helps for datasets that
big.
innodb_flush_log_at_trx_commit = 1 is a big performance kill
Nope, it was locked on a single value for about 36 hours, until we restarted the
engine last night. Now it's running fine, and we're setting up a testbed to
evaluate
MySQL 5.6 and FreeBSD 9 (?) for replacing our current Solaris 10/MySQL 5.1.46
setup.
On 6/28/13 12:44 AM, walter harms wrote:
Avoid FreeBSD,
Unless they did some real big magic on the scheduler of 9.
Claudio
On Jun 28, 2013 6:12 PM, "Andy Wallace" wrote:
> Nope, it was locked on a single value for about 36 hours, until we
> restarted the
> engine last night. Now it's running fine, and we're setting up a testbed
> to ev
Hi guys,
Did you have any experience running MyLSQ or in my case MariaDB 5.5.31
on FreeBSD on top of zfs?
We are using Samsung Pro 840 SSD drives and experiencing temporary stalls.
Our workload very much skewed towards inserts into big InnoDB tables
(70-100Gig) the dataset overall 1.5T.
I have
hi,
does the value change at all like below ?
mysql> show global variables like 'timestamp';
+---++
| Variable_name | Value |
+---++
| timestamp | 1372404355 |
+---++
1 row in set (0.00 sec)
mysql> show global varia
