VMwareDisks agent trigger to clean up stale disk entries that appear as error state in Veritas Volume Manager (VxVM) vxdisk command

Article: 100034089
Last Published: 2017-05-25
Ratings: 0 0
Product(s): InfoScale & Storage Foundation

Problem

In a Cluster Server (VCS) cluster, the VMwareDisks agent attaches the disks to a single virtual machine at a time.This helps to enable vMotion and Distributed Resource Scheduler (DRS) capabilities to a virtual machine in a VMware environment.The agent, as part of the online/offline operation, attaches the disks to the virtual machine and detaches it from the virtual machine.While detaching, the agent leaves stale VxVM entries.
 

Error Message

The detached disks by VMwareDisks resource are seen in error state with vxdisk command.

Solution

To resolve this issue Veritas has developed a new resstatechange trigger to clean up the stale VxVM entries after the VMwareDisks resource offline operation.

Functionality :
The trigger is executed for the following state changes of the VMwareDisks resources.
  1. ONLINE -> OFFLINE
  2. FAULTED -> OFFLINE  

The resstatechange trigger does the following:
 
  1. Identify if the disk is managed by VxVM.
  2. If the disk is managed by VxVM, remove the disk from VxVM configuration.
  3. Remove the device from the operating system (OS) device tree.
  4. Scan the disks.
Follow below configuration steps to add the mentioned trigger to the resource:
  1. Download the attachment which contains the resstatechange file.
  2. On all the nodes in the cluster, copy the resstatechange trigger to the “/opt/VRTSvcs/bin/triggers/” directory.
  3. On any node, enter the following commands to enable the resstatechange trigger for all the VMwareDisks resource:
       /opt/VRTSvcs/bin/haconf -makerw
/opt/VRTSvcs/bin/hares -modify <vmware_disk_res1> TriggerResStateChange 1
/opt/VRTSvcs/bin/hares -modify <vmware_disk_res2> TriggerResStateChange 1
/opt/VRTSvcs/bin/hares -modify <vmware_disk_res3> TriggerResStateChange 1
……
…..
/opt/VRTSvcs/bin/hares -modify <vmware_disk_resn> TriggerResStateChange 1
/opt/VRTSvcs/bin/haconf -dump -makero

Resource information in main.cf after modifications :

        VMwareDisks vmware_disk_res1 (
                TriggerResStateChange = 1
                ESXDetails = {
                         " server101.veritas.com" = "root=XsLUiSPuKShUJoGMhMImJ" }
                DiskPaths = {
                         "6000C297-282d-8f20-d84d-ec235f7b34ce:[SAN1] vm1/vm1_7.vmdk" = "1:0",
                         "6000C296-913f-ff78-9e29-4c6ec521497c:[SAN300G] vm1/vm1_8.vmdk" = "1:1" }
                VirtualDiskMode = independent_persistent
                )



 
 

References

Etrack : 3916381

Was this content helpful?