percona server performance drop

[Zak]

I been trying to trouble shoot this problem for sometime now in  percona.

I been trying to migrate from Mysql5.6 to  percona, however in our testings our business app performs terrible when multi master circular replication is on. 

Our setup, 2 servers same specs, running vmsphere 5.5. VMs are created and CentOs 6.5 installed.

On the same hardware, same VM setups, we are running Mysql 5.6 com edition with GTID circular replication with no problems in performance.

With  Percona as soon as I start slave, the master server performs badly with our business app.

If I had to run the same business app on the slave machine, while its receiving updates, no problems, but as soon as I make it a Master

the performance drops.

While all the binary logging stuff are on, without replication turned on by using start slave; the app performs fine as some people suggested that

there might be a performance hit  with sync-binlogs etc 

Ive started a thread about this on dba.stack which has my config files etc posted.

http://dba.stackexchange.com/questions/71188/slow-writes-on-percona-mariadb-server-only-when-replication-is-running?noredirect=1#comment128062_71188

This is so far the only issue that is stopping us from migrating to percona, and its driving me nuts. 

Our servers are going through a vpn and is lagging 0 seconds behind master

Does anyone have an idea what may be the problem. 

any percona engineers willing to comment.

thanks

zak

[Laurynas Biveinis]

Zak -

Can you send pt-pmp and SHOW ENGINE INNODB STATUS for both MySQL 5.6
and Percona Server 5.6, identical setup?

> --
> You received this message because you are subscribed to the Google Groups
> "Percona Discussion" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to percona-discuss...@googlegroups.com.
> To post to this group, send email to percona-d...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Laurynas
www.percona.com

[Zakir Mahomedy]

Laurynas

I managed to fine the problem. 

Thread of pools seems to be the culprit. 

The reason it worked fine on mysql 5.6 is because it doesn't support thread of pools. 

When I enable thread of pools, the problem occurs. When I disable it, everything seems to be fine.

Today I setup two VMs using percona to test replication on the LAN vs VPN. and I noticed

that thread of pools was the only thing different, so I commented it out and everything ran smoothly on the LAN.

I then decided, to turn off thread of pools on machine "254" and turn it on on "251".

app performance was bad on 251 and good on 254. I repeated the test a few times with by turning on/off thread of pools

and were able to reproduce the problem.  

I am assuming its thread of pools that may be causing this problem. We will need to test the solution over the vpn on a full load to 

be 100% sure that everything is functioning as it should be.

anyways here are the files as requested

we also noticed this problem when using tungsten replication  during our testing 

Zak

--
You received this message because you are subscribed to a topic in the Google Groups "Percona Discussion" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/percona-discussion/AlobUgU4dhM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to percona-discuss...@googlegroups.com.

[Laurynas Biveinis]

Ah. Your my.cnf has thread_pool_size = 2 which seems too low and
possibly explains the poor performance of thread pool. If you want to
use it, make sure to tune its settings,
http://www.percona.com/doc/percona-server/5.6/performance/threadpool.html