we setup drbd with ocfs2 on SSD Harddrives, to get the maximum performance for our webserver(s) and SQL Server.
Unfortunatly, the response time from the webserver is multiplicated with the factor 5-8 for ocfs2.
Without OCFS2, the response time is ~200 MS, with ocfs2 its about 1.5 seconds.
We tested drbd +ocfs2 in a primary/secondary configuration, no changes are made to the performance. Still the bad response time.
Than, we format the drbd drive with ext4 and get maximum performance, so the issue is related to ocfs2.
We followed the instructions on 2 different online howto´s , to setup ocfs2 (doesnt look heavy to do something wrong :)
is this normal, or are there "switches" to pull, to gain less response time.
The funny thing is, we have another cluster with drbd and ocfs2 running in production, with NFS on the DRBD + oxfs2 filer. The responsetime is very quick, but only if we use the nfs mount.
If we include a webserver node, with "direct" ocfs2 mount, we get also bad performance, so a lil hint would make us happy :)
thanks for your time
_______________________________________________ Ocfs2-users mailing list Ocfs2-users@oss.oracle.com https://oss.oracle.com/mailman/listinfo/ocfs2-users
