Re: [zfs-discuss] Is dedupe ready for prime time?
- "Paul Choi" skrev: > Roy, > > Thanks for the info. Yeah, the bug you mentioned is pretty critical. > In > terms of SSDs, I have Intel X25-M for L2ARC and X25-E for ZIL. And the > host has 24G RAM. I'm just waiting for that "2010.03" release or > whatever we want to call it when it's released... IIRC the memory requirements for dedup is something like 150 bytes per block for the DDT, meaning about 1GB per 1TB space if it's all 128kB blocks. With smaller blocks, osol gets greedy. Zil is good for write performance, but remember, osol won't use more than top half of memory size for zil, so you can probably slice it up and use the rest for l2arc. Vennlige hilsener roy -- Roy Sigurd Karlsbakk (+47) 97542685 r...@karlsbakk.net http://blogg.karlsbakk.net/ -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er et elementært imperativ for alle pedagoger å unngå eksessiv anvendelse av idiomer med fremmed opprinnelse. I de fleste tilfeller eksisterer adekvate og relevante synonymer på norsk. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Is dedupe ready for prime time?
Roy, Thanks for the info. Yeah, the bug you mentioned is pretty critical. In terms of SSDs, I have Intel X25-M for L2ARC and X25-E for ZIL. And the host has 24G RAM. I'm just waiting for that "2010.03" release or whatever we want to call it when it's released... -Paul On 5/18/10 12:49 PM, Roy Sigurd Karlsbakk wrote: - "Paul Choi" skrev: I've been reading this list for a while, there's lots of discussion about b134 and deduplication. I see some stuff about snapshots not being destroyed, and maybe some recovery issues. What I'd like to know is, is ZFS with deduplication stable enough to use? No, currently ZFS dedup is not ready for production. There are several bugs are filed, and the most problematic ones are that the system can be rendered unusable for days in some situations. Also, if using dedup, plan well your memory and spend money on l2arc, since it _will_ require either massive amounts of RAM or some good SSDs for l2arc. Best regards roy -- Roy Sigurd Karlsbakk (+47) 97542685 r...@karlsbakk.net http://blogg.karlsbakk.net/ -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er et elementært imperativ for alle pedagoger å unngå eksessiv anvendelse av idiomer med fremmed opprinnelse. I de fleste tilfeller eksisterer adekvate og relevante synonymer på norsk. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] Is dedupe ready for prime time?
- "Paul Choi" skrev: > I've been reading this list for a while, there's lots of discussion > about b134 and deduplication. I see some stuff about snapshots not > being > destroyed, and maybe some recovery issues. What I'd like to know is, > is > ZFS with deduplication stable enough to use? No, currently ZFS dedup is not ready for production. There are several bugs are filed, and the most problematic ones are that the system can be rendered unusable for days in some situations. Also, if using dedup, plan well your memory and spend money on l2arc, since it _will_ require either massive amounts of RAM or some good SSDs for l2arc. Best regards roy -- Roy Sigurd Karlsbakk (+47) 97542685 r...@karlsbakk.net http://blogg.karlsbakk.net/ -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er et elementært imperativ for alle pedagoger å unngå eksessiv anvendelse av idiomer med fremmed opprinnelse. I de fleste tilfeller eksisterer adekvate og relevante synonymer på norsk. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] Is dedupe ready for prime time?
I've been reading this list for a while, there's lots of discussion about b134 and deduplication. I see some stuff about snapshots not being destroyed, and maybe some recovery issues. What I'd like to know is, is ZFS with deduplication stable enough to use? I have two NFS servers, each running OpenSolaris 2009.06 (111b), as datastores for VMWare ESX hosts. It works great right now, with ZIL offload and L2ARC SSDs. I still get occasional complaints from developers saying the storage is slow - which I'm guessing is that read latency is not stellar on a shared storage. Write latency is probably not an issue due to the ZIL offload. I'm guessing deduplication would solve a lot of this read latency problem, having to do fewer read IOs. But is it stable? Can I do nightly recursive snapshots and periodically destroy old snapshots without worrying about a dozen VMs suddenly losing their datastore? I'd love to hear from your experience. Thanks, -Paul Choi ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
