On Thu, Nov 24, 2011 at 2:26 PM, Matthew Seaman <m.sea...@infracaninophile.co.uk> wrote: > On 24/11/2011 19:19, APseudoUtopia wrote: >> On Tue, Nov 22, 2011 at 3:06 AM, Matthew Seaman >> <m.sea...@infracaninophile.co.uk> wrote: >>> On 22/11/2011 02:09, APseudoUtopia wrote: >>>> Another quick question about swap: If I have 4 drives, with 512MB >>>> swap, the system uses all 4 swap partitions, correct? So it's not like >>>> it'd be going to waste? I'd have a total of 2 GB swap? >>> >>> Well, yes. If you just declare those raw partitions to be swap areas, >>> that will be the case. However, doing this is asking for trouble: you >>> subvert any resilience features obtained by using ZFS with raidz1. If >>> any one of the drives fails, your swap area will break and your system >>> will probably crash. >>> >>> Better to set up two pairs of gmirrors for swap -- the procedure is >>> described here: http://wiki.freebsd.org/RootOnZFS/GPTZFSBoot/Mirror >>> in section3 "Finish Install." This will effectively give you a raid10 >>> for your swap, with a total size of 1GB. >>> >> >> I'm not sure I understand this. How would that negatively affect the >> raidz1? The swap isn't in the zpool. I understand the system may crash >> if the OS was using the swap space and the drive failed. But would you >> not be able to reboot into a degraded zpool state and still have a >> usable system? >> > > No -- it means a failed disk can cause your system to crash. That's not > resilient behaviour. Yes, the data on the ZFS raidz1 should survive the > crash and the reboot, but the point is ZFS raidz1 should be able to > survive a disk failure like that /without/ a system crash. >
Ah! I understand. Thank you for the explanation. _______________________________________________ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
