TapeshGoyal
Contributor
Contributor

Esx sees deleted RDM lun paths

Hi ,

I have created a RDM lun on windows 2k3 GOS, esx 3.02, using FCP emulex hba and again deleted.

RDM got deleted from GOS but ESX is still seeing it.

esxcfg-mpath -l shows these deleted LUN.

Disk vmhba5:3:6 (3074MB) has 1 paths and policy of Fixed

FC 8:1.0 10000000c948cb98<->500a09818534c06e vmhba5:3:6 On active preferred

Disk vmhba5:6:6 (3074MB) has 1 paths and policy of Fixed

FC 8:1.0 10000000c948cb98<->500a09898534c06e vmhba5:6:6 On active preferred

But esxcfg-vmhbadevs does not show device.

vmhba0:128:0 /dev/sda

vmhba5:3:0 /dev/sdb

vmhba5:3:1 /dev/sdc

vmhba5:3:2 /dev/sdd

vmhba5:3:3 /dev/sde

vmhba5:3:4 /dev/sdf

vmhba5:6:4

vmhba5:3:5 /dev/sdg

vmhba5:3:6

vmhba5:6:6

Even rescan doesnt remove the paths.

vmkwarning shows these errors:

Jul 18 23:39:20 esx-225-150 vmkernel: 0:05:26:49.473 cpu1:1036)WARNING: SCSI: 6325: Failed for vmhba5:3:6: I/O error

Jul 18 23:39:20 esx-225-150 vmkernel: 0:05:26:49.491 cpu1:1036)WARNING: SCSI: 7917: status I/O error, rstatus #c0de00 for vmhba5:6:6. residual R 999, CR 80, ER 3

Jul 18 23:39:20 esx-225-150 vmkernel: 0:05:26:49.491 cpu1:1036)WARNING: SCSI: 6325: Failed for vmhba5:6:6: I/O error

Jul 18 23:39:21 esx-225-150 vmkernel: 0:05:26:50.329 cpu0:1042)WARNING: SCSI: 1736: Unexpected status returned: bad000a I/O error

Jul 18 23:39:21 esx-225-150 vmkernel: 0:05:26:50.347 cpu0:1042)WARNING: SCSI: 1736: Unexpected status returned: bad000a I/O error

Jul 18 23:39:28 esx-225-150 vmkernel: 0:05:26:57.656 cpu1:1035)WARNING: SCSI: 2215: Cannot remove target. World 1062 has not completely released the device.

Jul 18 23:39:28 esx-225-150 vmkernel: 0:05:26:57.656 cpu1:1035)WARNING: LinSCSI: 4591: The physical media represented by vmhba5:3:4 has changed and the device is in use. The device cannot be re-synchronized with the system. This is a critical error.

Jul 18 23:39:28 esx-225-150 vmkernel: 0:05:26:57.657 cpu1:1035)WARNING: SCSI: 2215: Cannot remove target. World 1062 has not completely released the device.

Jul 18 23:39:28 esx-225-150 vmkernel: 0:05:26:57.657 cpu1:1035)WARNING: LinSCSI: 4613: Device vmhba5:3:6 has disappeared but is currently in use and could not be removed.

Jul 18 23:39:36 esx-225-150 vmkernel: 0:05:27:04.849 cpu1:1035)WARNING: SCSI: 2215: Cannot remove target. World 1062 has not completely released the device.

Jul 18 23:39:36 esx-225-150 vmkernel: 0:05:27:04.849 cpu1:1035)WARNING: LinSCSI: 4591: The physical media represented by vmhba5:6:4 has changed and the device is in use. The device cannot be re-synchronized with the system. This is a critical error.

Jul 18 23:39:36 esx-225-150 vmkernel: 0:05:27:04.851 cpu1:1035)WARNING: SCSI: 2215: Cannot remove target. World 1062 has not completely released the device.

Jul 18 23:39:36 esx-225-150 vmkernel: 0:05:27:04.851 cpu1:1035)WARNING: LinSCSI: 4613: Device vmhba5:6:6 has disappeared but is currently in use and could not be removed.

Jul 18 23:39:38 esx-225-150 vmkernel: 0:05:27:07.540 cpu1:1035)WARNING: SCSI: 7917: status I/O error, rstatus #c0de00 for vmhba5:3:6. residual R 999, CR 80, ER 3

hostd dont show any errors.

Any idea why esx still sees this lun ?

Thanks,

Tapesh.

Tags (2)
0 Kudos
8 Replies
weinstein5
Immortal
Immortal

You have to rememebr what an RDM really is - it is a special VMDK that redirects the vmkernel to the LUN you want to write - by removing the RDM you simply remove the pointer to the LUN - your ESX will still be able to see the LUN because it is a LUN that is exposed to the ESX host from the SAN -

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
0 Kudos
TapeshGoyal
Contributor
Contributor

Yeah weinstein,

I understand that RDM is just present LUN to guestOS from ESX server.

My problem is even I have deleted the LUN from SAN and deleted RDM from

GOS, still ESX is able to see it.

The path to that LUN is hanging some where, and after rescan also it is

not removing.

Any clue for this would be appriciated.

Thanks,

Tapesh.

0 Kudos
weinstein5
Immortal
Immortal

OK I understand - have you tried rescanning from Service COnsole command line using the esxcfg-rescan command - also have you tried restart the vmware management agent using the command service vmware-mgmt restart to resynchronize with vc

If you find this or any other answer useful please consider awarding points by marking the answer correct or helpful
0 Kudos
depping
Leadership
Leadership

Just rescan the HBA's. ESX doesn't do this on automatically, so you will need to manually do this via the VI Client or the console -&gt; esxcfg-rescan vmhba1

Duncan

My virtualisation blog:

If you find this information useful, please award points for "correct" or "helpful".

0 Kudos
TapeshGoyal
Contributor
Contributor

Yeah tried rescan from both service console and VI-client, but didnt work. Even switching off VM also didnt worked. Then I reboot the ESX server and link removed. But rebooting is not always an option.

Any clue for this ?

Thanks,

Tapesh.

0 Kudos
Antonio_SP__Bra
Contributor
Contributor

Hi there,

did you delete the LUN with the virtual machine turned on? I'm asking because sometimes ESX gets confused when the contact with some LUN is lost.

Regards

Antonio

0 Kudos
kastlr
Expert
Expert

Hi,

how did you remove the LUN on your array?

In ealier SAN days, the HBA drivers didn't handle a RSCN (Request State Change Notification) from the SAN switch properly.

These RSCN did inform the HBA that something has changed in the environment.

The expected behaivior from the HBA would be to re-login into the fabric.

In combination with an old SAN switch firmware it could happen that a removed zone is still active untill the host reboots.

And some storage arrays does require some kind of front end activate/refresh to finally make changes active.

Hope this helps a bit.

Hope this helps a bit.      Greetings from Germany. (CET)
0 Kudos
ADMRC
Contributor
Contributor

Have you managed to sort this without a shutdown?

I am having the same issue.

0 Kudos