Hey Andy, On 10/31/2016 11:40 AM, Andy D'Arcy Jewell wrote: > We have been battling with this issue with help from IBM support, > but I'd like to understand the situation a bit better, so thought > I'd go to the source ;-) Thanks for reaching out; you've found someone. ;)
cat /etc/issue IBM_PowerKVM release 2.1.1 build 62 service (pkvm2_1_1) Kernel \r on a \m (\l)
The multipath-tools source code in PowerKVM 2.1.1 is older than what the upstream community works with -- this problem probably doesn't occur w/ the upstream sources. So, let me help you off list; I'll follow up w/ you privately.
So my questions are: a) Is this expected behaviour or a bug?
No, data corruption and WWID confusion are not expected; although the correspondence between bindings/wwids in initramfs and rootfs has been required on multipath-tools/distros for some time now. If you're right about no reboots (so the difference between wwids/ bindings in initramfs & rootfs is not a problem), then this is a bug.
b) If a bug, is there a fix?
Once we confirm the initramfs/rootfs mismatch is not a concern, and this is reproducible, yes, I'll be happy to land you a fix.
c) Is there any further information you need to help diagnose?
Yup. 1) Is this problem reproducible? ie, you can recreate it with a some steps in house? 2) If so, please enable verbosity 3 in multipath.conf, restart multipathd, and try to reproduce the problem. Then let's check the logs. I'll reply to your e-mail so we can follow up. cheers, -- Mauricio Faria de Oliveira IBM Linux Technology Center -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
