Hi,
I have a vm host which is crashing or has crashing vms when a zfs scrub is started. The disks of the host are somewhat older and not that fast.
As a resolution I changed the zfs module option zfs_vdev_scrub_max_active to 1 in /etc/modprobe.de/zfs.conf to reduce the number of scrub workers from 2(=default) to 1 to reduce the load imposed by zpool scrub.
My zfs.conf looks like this:
After changing that setting, the problem is gone.
I have a vm host which is crashing or has crashing vms when a zfs scrub is started. The disks of the host are somewhat older and not that fast.
As a resolution I changed the zfs module option zfs_vdev_scrub_max_active to 1 in /etc/modprobe.de/zfs.conf to reduce the number of scrub workers from 2(=default) to 1 to reduce the load imposed by zpool scrub.
My zfs.conf looks like this:
Code:
# 1 GB RAM for ZFS per 1 TB userdata capacity => ~ 6 TB Data
options zfs zfs_arc_max=6442450944 zfs_vdev_scrub_max_active=1After changing that setting, the problem is gone.
Last edited: