Unsolved
This post is more than 5 years old
9 Posts
0
67307
June 25th, 2012 08:00
Lost LUN on VM Windows 2008 on EQL
Hi.
I got 2 members on our EQL setup with loadbalancing, running 5.2.1.
I got 6 ESX hosts, connected to EQL.
1 VM (Win2008 with SQL 2008) got 3 LUNs directly mapped to EQL through a vswitch on vmware with one NIC (vmxnet2 enhanced)
3 times, one of the LUNs are going offline during high IO (backups), the other two LUNs are working fine.
I have talked to EQL support, and they can´t see any strange in the logs. EQL are loadbalancing the nics, and volume are gone after this loadbalancing.
Got this in eventlog:
iscsiprt:
Initiator received an asynchronous logout message. The Target name is given in the dump data.
Initiator could not find a match for the initiator task tag in the received PDU. Dump data contains the entire iSCSI header.
A connection to the target was lost, but Initiator successfully reconnected to the target. Dump data contains the target name.
mpio:
All paths have failed. \Device\MPIODisk1 will be removed.
A fail-over on \Device\MPIODisk1 occurred.
The device, \Device\Harddisk2\DR2, is not ready for access yet.
What tuning/changes could I do to VM that maybe solves this?
Is it any hotfixes from MS we should apply ?
Thanks, Magnus
DELL-Joe S
7 Technologist
•
729 Posts
0
June 25th, 2012 12:00
If you haven’t done so already, ensure the iSCSI disk timeout values for the W2K8 VM is setup per the “iSCSI Initiator and Operating System Considerations” document located on the FW download page. The VM will need to be restarted for the changes to take effect.
Ensure ESX is patched to the latest build (you didn’t specify what version you are running).
For the MPIO on the W2K8 VM: Is this the native MPIO or are you using the EqualLogic HIT? If using the EQL HIT, ensure you have the latest version installed as well.
For ESX 5.x you may want to consider setting the Login timeout value higher. We recommend setting this to 60 seconds, but it requires ESX patch ESXI500-201112001 to make the change.
Try newer (or a previous) NIC/HBA drivers and settings, you didn’t mention what you have now.
Consider updating the firmware to 5.2.4-H1
-joe
DELL-Joe S
7 Technologist
•
729 Posts
0
June 26th, 2012 13:00
If you have only one NIC setup on the ESX host for the vSwitch then you aren't running any type of multipathing.
For ESX, you may want to add a second NIC to the vSwitch, then use the EqualLogic MEM (Multipathing Extension Module for VMware) program to create the multipath on the ESX side.
For the VM, you can uninstall the optional MPIO component, just run the installation program and select the EqualLogic MPIO DSM component, you are prompted to restart the computer/VM.
-joe
magnus.tengmo
9 Posts
0
June 26th, 2012 13:00
Thanks for repsonse.
I verified registry settings and they are fine.
I´m was using vmxnet2 enhanced for iscsi, changed to vmxnet3 today.
Added a hotfix that MS support recommends: KB2522766
I used HIT 3,4 with MPIO, but upgraded to 4.0 now.
Running ESXi 4.1 with latest patches from this year. Build 702113.
Is it supported to have MPIO installed with only one NIC ?
Prerequisites for Configuring Multipath I/O DSM
Two or more supported GigaBit Ethernet (or faster) network interface cards (NICs) or host bus adapters (HBAs). See the Release Notes for more information.
Could I still use Autosnapshots manager, and not using EQL MPIO ? Is it possible to uninstall just the MPIO plugin?
Thanks, Magnus
dion.thanos
1 Message
0
June 26th, 2012 14:00
We ran into the same issue with a windows 2008 server and applied MS KB2522766.
I am pretty sure it works for both server 2008 and server 2008 R2. we had the exact same thing in the logs. Got the HF suggestion from TechSupport and the HF requires a reboot.
favelar
1 Message
0
August 6th, 2013 10:00
This used to happen in your environment with 2008 and 2008R2. The MS KB2522766 fixed the issue. Now we have moved on to Server 2012. and the issue has resurfaced. Has anybody seen this and found a fix for it?