I'm having an problem using mdadm to multipath two scsi devices. We are running this type of configuration on all our RedHat AS3 boxes with no issues but I have noticed that the mdadm on RedHat is a bit different than the one on gentoo. I've posted this with no luck on some gentoo forums. Below is the details of what is going on. If you could let me know if this is an issue of mdadm or possibly a gentoo issues and if you have any suggestions of a fix I would appreciate it. Thanks.
Nolan
I created the multipath device with the following command:
Code:
mdadm --create /dev/md0 --level multipath --raid-devices=2 /dev/sdc1 /dev/sdd1
This works fine. I can fail either of my devices and all is well.
I stop md0 and then start it again with this command:
Code:
mdadm --build /dev/md0 --level=multipath --raid-devices=2 /dev/sdc1 /dev/sdd1
and it still will fail over fine. When I look at /proc/mdstat all looks well:
Code:
Personalities : [linear] [multipath]
md0 : active multipath sdd1[1] sdc1[0]
97667136 blocks [2/2] [UU]
unused devices: <none>
Here is what shows up in my syslog:
Code:
Apr 7 11:42:38 dbinstall kernel: md: bind<sdc1>
Apr 7 11:42:38 dbinstall kernel: md: nonpersistent superblock ...
Apr 7 11:42:38 dbinstall kernel: md: bind<sdd1>
Apr 7 11:42:38 dbinstall kernel: md: nonpersistent superblock ...
Apr 7 11:42:38 dbinstall kernel: multipath: array md0 active with 2 out of 2 IO paths
I get then get the UUID of the array and write my mdadm.conf
Code:
DEVICE /dev/sd*
ARRAY /dev/md0 UUID=b6e63e9b:d4cbebcc:8144a88e:5dd85696
PROGRAM /usr/sbin/handle-mdadm-events
If I try and take advantage of the mdadm.conf file and execute this command to start md0:
Code:
mdadm -As
I get the following message:
Code:
mdadm: failed to add /dev/sdd1 to /dev/md0: Device or resource busy
mdadm: /dev/md0 has been started with 1 drive (out of 2).
When I take a look at /proc/mdstat here is what I get:
Code:
Personalities : [linear] [multipath]
md0 : active multipath sdc1[0]
97667072 blocks [1/1] [U]
unused devices: <none>
Here is what shows up in my syslog:
Code:
Apr 7 11:43:43 dbinstall kernel: md: md0 stopped.
Apr 7 11:43:43 dbinstall kernel: md: bind<sdc1>
Apr 7 11:43:43 dbinstall kernel: md: export_rdev(sdd1)
Apr 7 11:43:43 dbinstall kernel: multipath: array md0 active with 1 out of 2 IO paths
Only sdc1 shows up so when I test failover it doesn't work.
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
