On a reasonably up to date ZFS, if you "properly" (not just physically) remove it, the pool will just deal with it. Have tested this a couple of days ago. Make sure your ZFS version supports slog removal.
On Thu, Jan 22, 2015 at 3:32 PM, Nicholas Lee <[email protected]> wrote: > > > On 22 January 2015 at 10:54, Greg Zartman via smartos-discuss < > [email protected]> wrote: > >> was your disk marked as bad (and the pool marked degraded) or was there >>> some other way of determining your disk was "bad" ? >>> and what does your original test look like (was 90-ish MB/s with "bad" >>> disk plus slog) without the "bad" disk? >>> >> >> >> No, it looked fine when I did zpool status. I was just getting horrible >> performance in my KVM zone. I added the an intel 3700 and it didn't make >> much difference. SIGXCPU from the IRC channel helped me run an i/o test on >> all devices in the pool and the bad device was pegged at 100% usage while >> the slog and other devices where just idling. I pulled that one device >> from the pool and the performance in the KVM VM container shot up and the >> iostate of the devices showed pretty equal usage across the pool. >> >> > Was this bad disk that root cause of your KVM i/o issues? I guess it's a > bit hard to remove the slog now. > > Although wont the pool keep operating if you offline it? > > > Nicholas > ------------------------------------------- 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
