Hi Daniel I would recommend the 100G Hitachi SAS SSD SSD800MH.B - I’ve bought them for $750 AUD and they fly. I’ve seen sequential writes of 476MB/sec and 512K randoms at 313.5MB/sec.
- Dave > On 14 Apr 2016, at 6:56 PM, Dirk Steinberg <[email protected]> wrote: > >> >> Am 14.04.2016 um 02:39 schrieb Richard Elling >> <[email protected] <mailto:[email protected]>>: >> >> >>> On Apr 13, 2016, at 4:40 PM, Daniel Carosone <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>> Yes, agreed and understood. It is a space reservation that ensures some >>> number of blocks will never be allocated. >>> >>> That's not exactly the same as them never being used, due to CoW updates, >>> but it's very close. Once the pool is close to full, any writes that don't >>> immediately free the original blocks will get denied. >>> >>> The net effect is the same: a relatively constant number of free blocks for >>> the ssd controller to use in its own wear levelling and performance >>> management. Overprovisioned storage with lots of spare blocks above >>> whatever the device keeps internally already. >>> >>> At least, it seems so to me. My question, elaborated thus, is: what is the >>> difference you see that makes it insufficient? >>> >>> Oh, are we not issuing TRIM from zfs as space is freed? >>> >> no >>> That would explain it. If so, writing zeros into the reserved space >>> (without compression, dedup, or snapshots) occasionally will tell the ssd >>> controller the blocks are empty. >>> >>> I feel this is an effective workaround entirely within zfs, without >>> resorting to the ugly tricks of multiple partitioning schemes and >>> inflexible external allocations we both dislike. >>> >>> >> >> pedantic question: why not buy good quality SSDs? > > Hmm, price? My 2TB 850 EVO cost me 530 EUR. > How much would a „high quality“ SSD (say from Intel) cost? Maybe 2000 EUR? > > Also, availability in certain form factors (M.2) and capacities (I have never > seen > one of those HQ SSDs in 2 TB listed in a shop). > >> In my studies, good quality SSDs with >> decent overprovisioning perform more consistently than el-cheapos. > > That is certainly true. > >> FWIW, the preponderance of the evidence suggests that wear out is not as >> important as age. >> COW file systems like ZFS are particularly well behaved. >> https://www.usenix.org/conference/fast16/technical-sessions/presentation/schroeder >> >> <https://www.usenix.org/conference/fast16/technical-sessions/presentation/schroeder> >> https://users.ece.cmu.edu/~omutlu/pub/flash-memory-failures-in-the-field-at-facebook_sigmetrics15.pdf >> >> <https://users.ece.cmu.edu/~omutlu/pub/flash-memory-failures-in-the-field-at-facebook_sigmetrics15.pdf> >> >> -- richard >> >>> On 13 Apr 2016 18:27, "Dirk Steinberg" <[email protected] >>> <mailto:[email protected]>> wrote: >>> Am 13.04.2016 um 09:53 schrieb Daniel Carosone <[email protected] >>> <mailto:[email protected]>>: >>>> What is wrong with a dataset with refreserv set? >>>> >>> It does not actually reserve any specific blocks on the disk (LBAs for >>> SATA) which would >>> allow the SSD controller to deduct that a certain part of the SSD is not >>> being used. >>> >>> freservation is purely a (virtual) space accounting method of ZFS. >>> >> > > smartos-discuss | Archives > <https://www.listbox.com/member/archive/184463/=now> > <https://www.listbox.com/member/archive/rss/184463/27861131-64318a9b> | > Modify <https://www.listbox.com/member/?&;> Your Subscription > <http://www.listbox.com/> ------------------------------------------- smartos-discuss Archives: https://www.listbox.com/member/archive/184463/=now RSS Feed: https://www.listbox.com/member/archive/rss/184463/25769125-55cfbc00 Modify Your Subscription: https://www.listbox.com/member/?member_id=25769125&id_secret=25769125-7688e9fb Powered by Listbox: http://www.listbox.com
