On 25/01/2018 15:39, Reginald Beardsley via openindiana-discuss wrote:
I lost the disk on my OI_151a8 system. The swap area was going bad. So I booted
from a 2017.10 Hipster DVD, imported the old drive and copied everything to another
disk using a USB <-> SATA adaptor.
The evolving nature of ZFS makes the traditional miniroot problematic. If the
ZFS implementation in the miniroot image is older than the version on disk, it
I've been underwhelmed by the recovery from a failed disk on Solaris 10 u8,
however, I think much of the issue there was the HP Z400 BIOS. After migrating
from 1 TB to 2 TB disks by replacing individual disks, when I removed the last
1 TB disk it would not boot until I had swapped the appropriate cables. I have
a peculiar setup, 3 disks with the root pool in a 3 way mirror in s0 and a
RAIDZ1 export pool in s1. Resilvering the RAIDZ1 disks was *very* slow. It
would have been quicker to setup the root pool, create the export pool and then
zfs receive the export pool image from the mirrored scratch pool where I'd put
it before I started the migration.
Putting these into /etc/system (from an oi151a9 machine) and reboot
should shrink your resilver times (with an affordable IO slowdown for
set zfs:zfs_resilver_delay = 0
set zfs:zfs_scrub_delay = 0
set zfs:zfs_top_maxinflight = 127
set zfs:zfs_resilver_min_time_ms = 5000
Typical scrub/resilvers are an hour for 150 GB data, compared to
nearly a day with the default parameters (which let resilver/scrub
only run when nearly nothing runs on the machines).
You should also put these into /etc/system to get crash dumps
if something ever panics:
* disable multithread dump (hangs)
Hope that helps.
Dr.Udo Grabowski Inst.f.Meteorology & Climate Research IMK-ASF-SAT
KIT - Karlsruhe Institute of Technology http://www.kit.edu
Postfach 3640,76021 Karlsruhe,Germany T:(+49)721 608-26026 F:-926026
openindiana-discuss mailing list