Joshua D. Drake wrote: > On Wed, 2009-02-25 at 09:21 +0900, Jordan Tomkinson wrote: >> On Wed, Feb 25, 2009 at 12:05 AM, Aidan Van Dyk <aidan@xxxxxxxxxxx> >> wrote: >> * Greg Smith <gsmith@xxxxxxxxxxxxx> [090201 00:00]: >> > Shouldn't someone have ranted about RAID-5 by this point in >> the thread? >> You mean someone's actually still using RAID-5? >> ;-) >> >> What exactly is wrong with RAID5 and what should we have gone with? On top of the stuff Joshua wrote, there's also the "RAID 5 Write Hole". Quoting Wikipedia: "In the event of a system failure while there are active writes, the parity of a stripe may become inconsistent with the data. If this is not detected and repaired before a disk or block fails, data loss may ensue as incorrect parity will be used to reconstruct the missing block in that stripe. This potential vulnerability is sometimes known as the write hole. Battery-backed cache and similar techniques are commonly used to reduce the window of opportunity for this to occur." And in more detail from http://blogs.sun.com/bonwick/entry/raid_z "RAID-5 write hole... What's worse, it will do so silently -- it has no idea that it's giving you corrupt data." I sometimes wonder if postgres should refuse to start up on RAID-5 in the same way it does on VFAT or running root. :-) > RAID5 outside of RAID 0 is the worst possible RAID level to run with a > database. (of the commonly used raid level's that is). > > It is very, very slow on random writes which is what databases do. > Switch to RAID 10. > > Sincerely, > > Joshua D. Drkae > > >> -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
