On Sat, 12 Sep 2009, Doug Hughes wrote: > [email protected] wrote: >>> >>> >>> adding standard DRAM onto this with a super capacity for flushing the >>> ram to nonvolatile flash gives you the best of all worlds. Its still >>> relatively new and relatively expensive. I know of only one vendor using >>> this in mainstream, and that's Sun with their Readzilla and Logzilla >>> accelerators for FishWorks. (Other vendors may be doing similar, but I'm >>> not aware of them) >> >> although unless the drive includes battery backup for the DRAM you will >> loose a lot of data if the drive looses power. > oops. Typo on my part. super 'capacitor', which acts as a battery in that it > flushes the DRAM to nonvolatile flash storage
writing to flash takes a considerable amount of power, do they really put capacitors that large on the boards? (I sure haven't seen anything that looked like a capacitor that large in any of the disassembled pictures of flash drives that I see in the reviews) >>> As I mentioned earlier, using Flash for the ZFS intent log GREATLY >>> speeds up zfs with NFS over spinning disk media, even when striped. You >>> do pay some penalty in serial write bandwidth however. You become >>> bottlenecked on the aggregate write throughput of the flash devices. On >>> the other hand, if you're just pumping out TB of sequential data, >>> there's no reason to even consider flash in the mix. Spinning media >>> kicks butt in this arena (even tape does very well) >> >> are you comparing this to having the ZFS intent log on the same drives as >> your other data? or are you comparing it to having your ZFS intent log on a >> dedicated high-speed rotating drive? >> > zfs normally stripes the intent log across the data disks unless you tell it > otherwise (e.g. 46 disks on a Sun x45000 or x4540) > You can also put the ZIL on a dedicated high speed rotating drive. The x25E > outperforms both except in the case where you want high sequential > throughput. well given that the X25E is faster than the rotating storage I would expect it to be a win there as well. but the point I was trying to make was that I suspect most of the benifit is in the step from a striped intent log to an intent log on dedicated media, with the move from there to an X25 helping, but not as much. >>> The Fusion IO card is the fastest thing out there at the moment. It's >>> part flash, part DRAM, like the logzilla but comes in a PCI-E form >>> factor that you have to plug into a machine, and it's also very spendy, >>> but if you really want to speed up your DB by orders of magnitude, you >>> could shove some of these into a machine with a bunch of PCIE2 slots and >>> mirror them together and throw indexes or small tables on here. >> >> the Fusion IO card is fast, but it's also extrememly expensive. >> > aye.. I did mention that in the second sentence.. >>> So, I guess to summarize, before saying that flash writes aren't fast, >>> have a test drive of the X-25E. Flash got a deservedly bad wrap in the >>> past for write speed, but things have come a long way in the last couple >>> of years. >> >> if you really need fast writes you need battery backed ram. if you don't >> need very much (up to a couple of gigs with the writes being converted from >> many small writes to a few large writes by the hardware) you can use a raid >> card/chassis with it. if you need a lot you can look at something like >> ACard's ANS-9010 Serial ATA RAM disk ( >> http://techreport.com/articles.x/16255 ), it's as much faster than a X25E >> as an X25E is from a normal hard drive. >> > I have evaluated a battery backed ram that was much slower than flash for ZIL > purposes. I will spare the vendor some embarassment, but I was very surprised > with the results and still don't fully understand why. no question that you can find lemons in any technology, many raid cards have traditionally just tried to be faster than the drives that they ran (enough to show an improvement), but the good ones compete with each other on performance, so with any of them I would not expect that sort of problem. David Lang _______________________________________________ Discuss mailing list [email protected] http://lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
