Hi,
I'm running Opensolaris 2009.06, and I'm facing a serious performance loss with
ZFS ! It's a raidz1 pool, made of 4 x 1TB SATA disks :
zfs_raid ONLINE 0 0 0
raidz1 ONLINE 0 0 0
c7t2d0 ONLINE 0 0 0
c7t3d0 ONLINE 0 0 0
c7t4d0 ONLINE 0 0 0
c7t5d0 ONLINE 0 0 0
In the beginning, when the pool was just created (and empty !), I had the
following performances :
- Read : 200 MB/s
- Write : 20 MB/s (10 MB/s with compression enabled)
These performances were OK at this time.
However, after 2 months of production use, and a volume of data of 1TB only,
the performances are near :
- Read : 5 MB/s
- Write : 500 KB/s !!!!
The write speed is so low that is breaks any network copy (Samba or SFTP). The
only solution I found to copy large files to the pool without outages is to use
SFTP via Filezilla, with the activation of the bandwich limit (limit=300 KB/s)
!!!
In this pool, I have 18 filesystems defined :
- 4 FS have a recordsize of 16KB (with a total of 100 GB of data)
- 14 FS have a recordsize of 128KB (with a total of 900GB of data)
There is a total of 284 snapshots on the pool, and compression is enabled.
There is 3 GB of physical RAM.
The usage of the pool is for daily backups, with rsync. Some big files are
updated simulteanously, in different FS. So, I suspect a huge fragmentation of
the files ! Or maybe..., a need of more RAM ??
Thank you for any thoughts !!
Philippe
--
This message posted from opensolaris.org
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
