1,225 reputation
39
bio website nex7.com
location Mercer Island, WA
age
visits member for 3 years, 5 months
seen Jul 8 at 18:28

( http://www.nexenta.com )

I am a Solutions Architect at Nexenta Systems, Inc, formerly a Nexenta Support Engineer.

My involvement in this website is strictly voluntary, and on my own initiative as time allows. Please do not message me privately about active NexentaStor support cases. If you have a Nexenta or ZFS-related question on ServerFault, I don't mind if you ping me with a link and request I take a look, but I cannot guarantee a response, much less a timely one.

I do not represent Nexenta in any official capacity on this site, and my comments/posts are my own, and do not represent official Nexenta responses or policy.


May
28
comment ZFS performance: do I need to keep free space in a pool or a file system?
Personally I use 70-75% maximum, not 85%. Edge cases can even degrade lower than that (I won't scare you with how much lower). You can in fact set a reservation to limit space.. this is common at some sites I know of. All you do is create a dataset, and set a reservation (or refreservation) on it that is equivalent to 25% of the total pool size, and poof, even if you use up every free block in your real datasets, you know the pool is still 25% free space, as you have a 25% reservation on there.
May
27
awarded  Editor
May
27
comment Tuning ZFS scrubbing, 141KB/s running for 15 days
Fair enough - added it all into the answer.
May
27
revised Tuning ZFS scrubbing, 141KB/s running for 15 days
per request, combining everything into answer
May
27
comment Tuning ZFS scrubbing, 141KB/s running for 15 days
This is pretty much completely false. CPU & RAM have effectively zero impact on scrub operations (assuming there's any free at all). Having lots of free RAM & CPU will not 'speed up' scrub operations. Scrub is limited by watching incoming I/O to the pool, not by checking for 'available system downtime', whatever that is.
May
10
answered ZFS iSCSI and free capacity in VMFS
May
9
comment Tuning ZFS scrubbing, 141KB/s running for 15 days
The %b and asvc_t reported imply some very, very random read workload going on (spinning disks should do better than that if it is truly sequential). First I'd turn on zfs_no_scrub_prefetch, to disable prefetch on scrub operations, just to see if that helped. If no joy, depending on the version of Nexenta you're on - you may be running 30/5, 5/1 or 10/5 (zfs_txg_timeout & (zfs_txg_synctime_ms*1000). Change zfs_txg_timeout to 10 and zfs_txg_synctime_ms to 5000, then try upping zfs_scan_min_time_ms to 3000 or 4000. This tells ZFS it can spend a lot longer on scrubs, may starve normal I/O!
May
9
comment Tuning ZFS scrubbing, 141KB/s running for 15 days
I should also read the entire question before responding - and stop browsing ServerFault while on conference calls. :)
May
9
comment Tuning ZFS scrubbing, 141KB/s running for 15 days
I suppose I should note that you modify these settings in bash using "echo <tunable>/W0t<number> | mdb -kw". And you view current values with "echo <tunable>/D | mdb -k". My notes say all of these can be changed in flight, none seem to require an /etc/system modification and reboot to take effect.
May
9
answered Tuning ZFS scrubbing, 141KB/s running for 15 days
Apr
15
awarded  Yearling
Apr
4
comment Are snapshots and clones filesystem-wide in ZFS-backed Lustre clusters
You've already answered your own question - no. I do have a question, though; what's your use-case, that you're after a distributed filesystem on Linux with CoW-like cloning and snapshots?
Mar
26
comment ZFS import pool ahcich0: timeout (NanoBSD 8.2 -> 9.1)
Does the pool still import fine on the old NanoBSD 8.2 USB stick? Because this just looks like a hardware error to me?
Mar
22
comment Accidently set chown “root” to all my server
I cannot vouch for this link, as I've never used it myself, but.. sysadminnotebook.blogspot.com/2012/06/…
Mar
22
answered Accidently set chown “root” to all my server
Mar
22
awarded  Critic
Mar
22
answered FreeNAS: Renaming a zpool
Mar
22
comment ZFS iSCSI and free capacity in VMFS
It would be around 100 GB, ASSUMING no snapshots. Throw snapshots in there, and that could quickly go well over 100 GB. Look into 'sdelete', btw. If you have compression on on the dataset (which you should), and no snapshots, sdelete will 'reclaim' most of the space on ZFS that isn't actually used by the filesystem on top through virtue of the zerofilling it does. Of course, this is an active task - you'd need to do it any time you wanted to reclaim space.
Mar
22
comment Are ZFS clustered filesystems possible?
Namespace Cluster has its uses - not sure this is one of them, though. I hesitate to answer this question -- for starters, the current forerunner answers (Gluster, Lustre, etc) have been referenced, and for seconds, my immediate response is not a how-to, but a 'why?'
Mar
22
comment Would storing logical volumes on a ZFS storage benefit from caching?
Careful. The default on illumos-derivatives utilizing COMSTAR (most of them) is actually usually to NOT use the ZIL for normal iSCSI traffic. You need to be sure 'writeback cache' is NOT enabled on the LU, or most setups will end up completely skipping ZIL mechanics.