Open jpoggi opened 3 weeks ago
If your pool got degraded, it's more probably a kernel problem than one of multipath-tools. Multipath-tools main responsibility is to reinstate paths that have been unplugged and re-plugged. If that didn't work, it might be a multipath-tools issue.
You need to provide your system logs for us to judge. Please use an appropriate attachment format or some pastebin.
Thanks for the quick answer.
I will generate some logs later next week I hope by unplugging the SAS cable.
Happy Halloween !
Hi,
I have the following configuration.
Operating system running on a
Dell PowerEdge R660
All the drives are running on a
Western Digital Ultrastar Data60
link with two SAS cables.Version of multipath-tools:
Here is a partially
multipath -ll
exportThe following specific configuration of multipath (I have already try path_grouping_policy failover with the same result) :
The following global configuration of multipath:
So the problem is the following, when I unplugged a SAS cable to test the redundancy of the multipath, I loose one path but my ZPOOL on top of it became
DEGRADED
and the only way to recover it is to hard reboot my server after having plugged my cable again of course.My guess is that I do not have the correct configuration for my Western Digital Ultrastar Data60, of I have a conflict between zfs on linux and multipath.
So any help would be appreciate :-)
Regards.