VMware Cloud Community
DD1953
Contributor
Contributor

iSCSI Datastore inactive

Hi all,

After updating ESX Host to VMware ESXi, 6.5.0, 13635690 my iSCSI Datastore is inactive and VM´s are inaccessable.
iSCSI Target is Synology RS812RP+

Any ideas?

Thanks in advance

Reply
0 Kudos
13 Replies
vXav
Expert
Expert

Is the Lun detected by your host in the devices view?

  • A) If yes is it attached or detached?
    • C) If detached can you attach it with right-click>Attach. If ok go to D)
    • D) Is datastore still inactive? If yes: try to mount the datastore with right-click>Mount
  • B) If not detected can you rescan your software adapter?
    • If ok > Go to A)
Reply
0 Kudos
DD1953
Contributor
Contributor

Hi vXav,

unfortunately it is E)

Its detected and attached, detaching not possible "device in use"

rescans not succesful

Reply
0 Kudos
vXav
Expert
Expert

Do you get the option to mount or unmount the datastore when right clicking on it?

Can you toggle that status whatever it is?

Also can you check that the MTU is still correct?

Reply
0 Kudos
DD1953
Contributor
Contributor

Hi,

I can`t mount or unmount, this options are greyed

Reply
0 Kudos
Raj1988
Enthusiast
Enthusiast

I do not see that model (Synology RS812RP+) in the supported list .

VMware Compatibility Guide - Storage/SAN Search

Can you rescan adapter and cehck the vobd.log and vmkernel.log on ESXi and copy over them.

Regards,

AJ

Reply
0 Kudos
DD1953
Contributor
Contributor

Hi AJ,

thanks for interest.

I´ll try to send logs tommorrow. We have a day off here today

BR

Reply
0 Kudos
DD1953
Contributor
Contributor

Hi AJ,

the logfiles

Reply
0 Kudos
vXav
Expert
Expert

And have you checked the MTU ? end to end

Reply
0 Kudos
Raj1988
Enthusiast
Enthusiast

vobd.log

2019-06-07T14:06:34.853Z: [iscsiCorrelator] 158390385us: [vob.iscsi.target.connect.error] vmhba65 @ vmk1 failed to login to iqn.2000-01.com.synology:VMNAS01.name because of a network connection failure.

2019-06-07T14:06:34.853Z: [iscsiCorrelator] 158390809us: [esx.problem.storage.iscsi.target.connect.error] Login to iSCSI target iqn.2000-01.com.synology:VMNAS01.name on vmhba65 @ vmk1 failed. The iSCSI initiator could not establish a network connection to the target.

2019-06-07T14:06:34.853Z: An event (esx.problem.storage.iscsi.target.connect.error) could not be sent immediately to hostd; queueing for retry.

2019-06-07T14:06:41.902Z: [vmfsCorrelator] 165439821us: [vob.vmfs.ats.support.lost] ATS-Only VMFS volume 'DS_VMNAS01' is not mounted. This host does not support ATS, or ATS initialization failed.

2019-06-07T14:06:41.902Z: [vmfsCorrelator] 165440253us: [esx.problem.vmfs.ats.support.lost] ATS-Only VMFS volume 'DS_VMNAS01' will not be mounted. Host does notsupport ATS or ATS intialization has failed.

Run below command on SSH to ESXi to check the locking mode.

esxcli storage vmfs lockmode list

You see output similar to:

Volume Name UUID Type Locking Mode ATS Compatible ATS Upgrade Modes ATS Incompatibility Reason

----------- ----------------------------------- ------ ------------ -------------- ----------------- ---------------------------

SYN01 56f52615-c7d247e4-e688-90b11c268865 VMFS-5 ATS true No upgrade needed

SYN02 578356c4-5abc30bc-2cbd-a4badb2db0d4 VMFS-5 ATS+SCSI false None Device does not support ATS

Unmount the VMFS datastore from all the ESXi hosts which are using it and keep it mounted to the ESXi on which its locking mode is showing as ATS.

Run the below command to set the locking mode to ATS+SCSI:-

esxcli storage vmfs lockmode set -s -l <VMFS_label>

Ex:- esxcli storage vmfs lockmode set -s -l DS_VMNAS01

After running the above command, all the ESXi hosts which were not able to see the datastore will be able to see it now.

Regards,

AJ

Reply
0 Kudos
DD1953
Contributor
Contributor

Hi AJ,

locking mode is set to ATS+SCSI

but datastore is still inactive

Reply
0 Kudos
DD1953
Contributor
Contributor

Hi vxav,

not sure what I should check? MTU size in TCP/IP?

It´s set to 1500, I think it was also 1500 before update.

BR

Reply
0 Kudos
vXav
Expert
Expert

When using iSCSI the MTU is usually set to 9000. Meaning you pass more data per frame to maximize bandwidth usage. It needs to be the same on all components of the chain (array > switch > nic).

However I think Ajay is onto something here.

I think he wants you to check the lockmode mecanism on all hosts to see if one of them has the datastore as ATS-only.

You can understand more about your issue in this KB.

Reply
0 Kudos
DD1953
Contributor
Contributor

Thank you guys,

I´ve bougth a support ticket and opened a case.

They check some logs now...

Reply
0 Kudos