Unsolved
1 Rookie
•
55 Posts
0
1862
virtual disk preferred path failed on MD3220i SAN
We are using this san last 4 year without any issue
But last 3 months ago one of raid control card lost and we replaced with
other one and it's working fine
But some reason we are getting "virtual disk preferred path failed " on
one of the virtual disk. So every time we have to set the prefered path manually
and its works fine for 15 days but after 15 days (during media scan)
it's back to another path with this error.
So could you please help me with this issue.
DellEMCSupport
Moderator
Moderator
•
631 Posts
0
April 20th, 2020 10:00
Hi,
Does it go through a switch? It could be a cabling issue.
Vijay Ayar
1 Rookie
1 Rookie
•
55 Posts
0
April 22nd, 2020 03:00
Yes, all cables are connected by a switch. And we are able to ping of all IP of raid controller. We are getting ping from all VM host machines too. So what could be the reason to fail the virtual path?
DiegoLopez
2.7K Posts
0
April 22nd, 2020 06:00
Hello @Vijay Ayar,
This error message is quite common. What happens is that the storage array automatically assigns a path for the I/Os. These paths are automatically designated according to load/speed and other factors. Sometimes it happens that the storage array cannot send this data by its preferred path and alerts us to this. This is not necessarily a problem, it can be a normal part of the storage array operation. In fact, thanks to this path modification it allows to continue to functioning.
Other times it is a problem, when the original channel is, for some reason, deteriorated. Generally, it has 3 possible points of failure: the storage array itself (degraded ports), the intermediate data switches (configurations, wiring..) or the configuration of the host to which the storage array is connected.
As far as the storage array is concerned, the first thing to do is to check that all the ports are correct and if there are VDs in the two slots of the controller.
Once we can say iy is not a problem of the storage array, it is either a problem of the intermediate switches (degraded ports, configuration?) or the configuration of the hosts. This happens usually with hosts that have VMWare ESXI. Do you have this OS? It is quite common with VMware that the configuration may not be correct. It happens that sometimes a fixed path is configured for a datastore instead of the Round Robin configuration. Please, in case you have VMWare ESXi check the configuration of the hosts and tell me which option is configured.
Therefore, before you do the operation of re-distributing the virtual disks (Which is indeed the solution to the error message). You must confirm that the original channel is not degraded. Otherwise, you will force the I/Os to a degraded path which will result in the data store being dropped.
Regards.
Vijay Ayar
1 Rookie
1 Rookie
•
55 Posts
0
April 23rd, 2020 03:00
Thank you so much for your reply.
So we are using the same SAN from the last few years without any changes/errors and it was working fine. We are using VMWare ESXi and SAN is attaching the same server via the switch.
Due to the raid controller issue we replaced the raid controller and after than we are facing the warning. F.Y.I Raid controller is working fine without any error.
DiegoLopez
2.7K Posts
0
April 23rd, 2020 06:00
Well, then for me it's clear than the Controller change also produced a configuration change.
Did you check if you have VDs on both controllers?
Did you check the VMware configuration? Make sure Round Robin is the active policy.
I would start with this two checks.
Regards.
Vijay Ayar
1 Rookie
1 Rookie
•
55 Posts
0
April 24th, 2020 02:00
We created one disk pool. Under one disk pool, We created 2 virtual disks
Like VD1 is running on Raid 0
and VD2 is running on Raid 1
We had set iscsi port IP on both raid controller as mentioned below...
Raid controller 0 : 10.10.6.212,10.10.7.212,10.10.8.212,10.10.9.212
Raid controller 1 : 10.10.6.111,10.10.7.111,10.10.7.111,10.10.7.111,
-----------------
So now Might be there is an issue going on our one of vm esxi server that,
Under VM ESXi 5.5 console..
Configuration -> Storage -> select VD1
under VD1 Properties
I have opened manage path (Round Robin) and found 10.10.6.212,10.10.7.212,10.10.8.212,10.10.9.212 running active(i/o)
Same way...
Configuration -> Storage -> select VD2
under VD2 Properties -
I have opened manage path (Round Robin) and found 10.10.6.212,10.10.7.212,10.10.8.212,10.10.9.212 running active(i/o)
So on both VD's, I am getting the same path in VM host .
So can you please tell the how to add other raid controller path into VD1
DELL-Sam L
Moderator
Moderator
•
6.8K Posts
0
April 24th, 2020 16:00
Hello Vijay Ayar,
What is the current version of firmware that you are running? How many iscsi connections per controller are you using? Is your MD3220i connect to a switch or directly to your host? Do you just have the one disk pool, or do you have more than one?
Please let us know if you have any other questions.
Vijay Ayar
1 Rookie
1 Rookie
•
55 Posts
0
April 29th, 2020 06:00
We have set all san cable connected to switch and also we have set all VM host cable to on the same switch.
So now i worked on VM host and found some iscsi path issue which I have corrected by rescanning iscsi adapter on vm host. So now I hope this solution can resolve my issue.