Has anyone seen this before?
I'm experiencing disk stalls while doing sysupgrade.
OpenBSD 6.7 -> 6.8, which is a test VM running on KVM RHEL7/RHEL8.
VM storage is LVM on top of 10Gbps iSCSI with multipath.
Disk and network are on virtio.
Installation is fine (iso image stored locally on node).
But when doing sysupgrade I experience disk stalls.
Also qemu-kvm process has high cpu usage.
strace on qemu-kvm process on Linux node shows a hell lot of:
14:36:29 ppoll([{fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=7,
events=POLLIN}, {fd=8, events=POLLIN}, {fd=13, events=POLLIN}, {fd=16,
events=POLLIN}, {fd=18, events=POLLIN}, {fd=29, events=POLLIN}, {fd=31,
events=POLLIN}, {fd=32, events=POLLIN}, {fd=33, events=POLLIN}, {fd=36,
events=POLLIN}], 12, {tv_sec=0, tv_nsec=996416}, NULL, 8) = 0 (Timeout)
<0.001011>
14:36:29 ppoll([{fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=7,
events=POLLIN}, {fd=8, events=POLLIN}, {fd=13, events=POLLIN}, {fd=16,
events=POLLIN}, {fd=18, events=POLLIN}, {fd=29, events=POLLIN}, {fd=31,
events=POLLIN}, {fd=32, events=POLLIN}, {fd=33, events=POLLIN}, {fd=36,
events=POLLIN}], 12, {tv_sec=0, tv_nsec=996427}, NULL, 8) = 0 (Timeout)
<0.001011>
If I logout from one of multipath the timeouts above stop, stalls stop and
upgrade continues.
If I don't logout, then upgrade will finish but it will take a lot ~ 10min.
I have a bugzilla ticket on redhat about this.
https://bugzilla.redhat.com/show_bug.cgi?id=1721395#c17
Rest of the infrastructure (what is not openbsd, is Linux) works fine.
I'm experiencing this for a long time (probably after sysupgrade was
introduced).
Manually upgrading, while system is running, does NOT stall.
G
ps. I have feeling that this might be related to kvm_intel preemption_timer
I've seen this before on this list, but it's not easy for me to test it right
now.
