Possibly weird behavior with mpath_persist on CentOS 6.6 (device-mapper-multipath-0.4.9...)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi folks,

We are seeing what looks like weird behavior with the mpath_persist
command from the CentOS 6.6 package
(device-mapper-multipath-0.4.9...).

For background we have 25 iSCSI targets, each with four LUNs and four paths.

This results in a total of 400 devices (/dev/sdxx).

When we do an mpath_persist to register a key on some
/dev/mapper/mpathNN device, we see some 420 reads of 8 sectors (4096
bytes) at LBA 0 in the packet capture. At other times all we see is
streams of NOP IN or NOP OUT requests.

It seems like, but we have not verified it yet, that the mpath_persist
has triggered a read on every device, perhaps to check if the paths
are all OK, but I would expect it to do so only on the four paths for
the actual device we are trying to register a key for.

Can someone suggest if I have the wrong idea here? We are reading the
code and checking other things to see what is going on, but I thought
I would check to see if someone else knows about this behavior.

-- 
Regards,
Richard Sharpe
(何以解憂?唯有杜康。--曹操)

--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel




[Index of Archives]     [DM Crypt]     [Fedora Desktop]     [ATA RAID]     [Fedora Marketing]     [Fedora Packaging]     [Fedora SELinux]     [Yosemite Discussion]     [KDE Users]     [Fedora Docs]

  Powered by Linux