[DRBD-user] Secondary node saturises RAID array

Joris van Rooij jorrizza at wasda.nl
Tue Apr 15 15:01:30 CEST 2008

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


> Thanks, and I'll report my findings back to this list.

The problem wasn't caused by the RAID-6 setup. On two RAID-10 arrays the 
problem persisted. The sync system calls still take quite long. Write 
performance was actually a bit lower, but nothing to be worried about.

MySQL commit performance gains of almost 1500% were achievable by using the 
following MySQL parameter: innodb_flush_log_at_trx_commit = 2. It simply 
limits the amount of calls to sync() to once a second. 

Of course this isn't an ideal situation, but it's good enough to hit 
production. If anyone has an idea about how to speed up sync(), that'll be 
really helpful, still.

Thanks for all the help sofar.


More information about the drbd-user mailing list