Unsolved
3 Posts
0
777
Disk Failure - PowerVault MD3820f
Hello dear colleagues!
We have a PowerVault MD3820f and 2 disk groups on it. Last night we have a disk failure in one of disk groups with these symptoms:
- At 1:18AM we received first alert - Event type: 100D, Event category: Error, Description: Physical disk command timed out
- Several messages like that in next 14 minutes
- At 1:32AM at last we've received event - Event type: 201E, Description: Virtual disk repair started
Out system is built on MD3820f & Brocade 6505 FC Switch & R440 HV Servers
Between 1:18AM and 1:32AM we have a loss of access from servers to Virtual Disk from Disk Group in which was the failed disk (our hypervisor has almost gone mad during these 14 minutes). As VD repair started - connectivity was restored.
How can we avoid such losses of connectivity in future? It seems strange for us to have such a long connectivity loss.
Controller firmware version: 08.25.14.60
Physical disks (all) - Seagate DL1800MM0159, HDD Firmware ST51
Update: yes, the disk has failed(
DELL-Sam L
Moderator
Moderator
•
6.9K Posts
0
May 22nd, 2019 07:00
Hello mbolsh,
How many connections from each controller to your switch do you have? Are you only using a single switch that connects your MD3820f to your hosts? What is the OS that is running on the hosts? What multipathing policy are you using as well?
Please let us know if you have any other questions.
mbolsh
3 Posts
0
May 23rd, 2019 00:00
There are two links from each controller. They are connected to two Brocade FC Switches. So every VD has four links to every host. Every host has two connections to FC switches.
OS - VMware 6.5. Multipathing policy - Most Recently Used.
DELL-Sam L
Moderator
Moderator
•
6.9K Posts
0
May 24th, 2019 06:00
Hello mbolsh,
I would change your multipathing to use round robin instead of most recently used. Are your virtual disk balanced so that both controllers own half, or are all virtual disk own by the same controller?
Please let us know if you have any other questions.