Loss of terminal connection whilst vxevac operation is in progress may result in vxplex ERROR V-5-1-809 Plex <name> in <volume>is locked by another utility & VxVM V-5-1-10128 Transaction aborted waiting for io drain

book

Article ID: 100051923

calendar_today

Updated On:

Description

Error Message

# vxplex -g datadg -f det %1
VxVM vxplex ERROR V-5-1-809 Plex %1 in volume sasvol is locked by another utility

 

Cause

When the vxevac operation is interrupted it can cause the temporary plex to go into a locked state.
This is to safeguard against plex corruption, due to the failed subdisk migration to the new plex.
 

How to clear the "locked by another utility"


The following article outlines the step to clear the "locked by another utility" state;
 https://docs.infoscale.com/support/en_US/article.100037758

The following commands should to be executed against the problematic plex first, followed by the associated volume name.
 
In this instance, the problematic plex that cannot be dissociated is "%1".

# vxmend -g datadg -r clear all %1

The related volume name is "sasvol".

# vxmend -g datadg -r clear all  sasvol

Important: Confirm using "vxprint -qhtg " that the problematic plex is not associated with any volume, prior to running the following plex clean-up operation.

To delete the stale subdisk records associated with the problematic plex "%1 (for which is not associated with any volume at this time), type:


# vxedit -g datadg -rf rm %1

This can results in the following error state:

# vxplex -g datadg -f det %1
VxVM vxplex ERROR V-5-1-10128 Transaction aborted waiting for io drain


Even flushing the disk group configuration to disk, does not help.
# vxdg flush

The following commands are unable to resolve the issue where CVM (Clustered Volume Manager) is enabled.

When attempting to switch the CVM Master role, the command hangs on the current CVM Master node:
# vxclustadm setmaster
 

Resolution

The "VxVM vxplex ERROR V-5-1-10128 Transaction aborted waiting for io drain" state can be fixed as follows:
 

1.] Stop VCS on the CVM Master node

# vxdctl -c mode

# hastop -local


2.] If VCS fails to stop cleanly. The CVM master server may need to be power-cycled

3.] Once the CVM Master role has been switch to an alternate Slave node, connect to the new Slave and run the following commands
 

On new CVM Master:
 

# vxprint -qhtg datadg
dg datadg   default      default  38000    1524849274.262.fred
 
dm datadg01      purestorage1_0001105a auto 65536 10737352400 -
dm datadg02      purestorage1_00011059 auto 65536 10737352400 -
dm datadg03      purestorage2_000119c7 auto 65536 9017687760 -
dm purestorage4_00011ab7 purestorage4_00011ab7 auto 65536 9017687760 -
sd purestorage4_00011ab7-01 %1 purestorage4_00011ab7 0 8367264848 21474704800 purestorage4_00011ab7 ENA

v  sasvol -           ENABLED  ACTIVE   29841969648 SELECT -        fsgen
pl sasvol-01 sasvol ENABLED ACTIVE 29841969648 CONCAT -      RW
sd datadg-01   sasvol-01 datadg01 0     10737352400 0      purestorage1_0001105a ENA
sd datadg-02   sasvol-01 datadg02 0     10737352400 10737352400 purestorage1_00011059 ENA
sd datadg-03   sasvol-01 datadg03 0     8367264848 21474704800 purestorage2_000119c7 ENA


NOTE: The TEMP plex "%1" has been automatically deleted when the CVM Master is switched to a new node.

4.] If this is the case, the orphaned subdisk (purestorage4_00011ab7-01) can be deleted using the following syntax:

# vxedit -g datadg -rf rm purestorage4_00011ab7-01

5.] If the disk in a thin LUN, the disk space must be reclaimed in order to free up the space currently used by the orphaned subdisk

Use the additional "-z" flag argument to list the subdisks to be reclaimed
 

# vxprint -qhtg datadg -z
dg datadg   default      default  38000    1524849274.262.fred
 
dm datadg01      purestorage1_0001105a auto 65536 10737352400 -
dm datadg02      purestorage1_00011059 auto 65536 10737352400 -
dm datadg03      purestorage2_000119c7 auto 65536 9017687760 -
dm purestorage4_00011ab7 purestorage4_00011ab7 auto 65536 9017687760 -
sd purestorage4_00011ab7-01 - purestorage4_00011ab7 0 8367264848 - purestorage4_00011ab7 RCLM

v  sasvol -           ENABLED  ACTIVE   29841969648 SELECT -        fsgen
pl sasvol-01 sasvol ENABLED ACTIVE 29841969648 CONCAT -      RW
sd datadg-01   sasvol-01 datadg01 0     10737352400 0      purestorage1_0001105a ENA
sd datadg-02   sasvol-01 datadg02 0     10737352400 10737352400 purestorage1_00011059 ENA
sd datadg-03   sasvol-01 datadg03 0     8367264848 21474704800 purestorage2_000119c7 ENA

 

6.] To reclaim the space against the specific disk, type:

# vxdisk -g datadgreclaim purestorage4_00011ab7 &

The progress can be monitored using:

# vxtask -l list

7.] Once the thin reclaim operation is complete, the vxevac operation can be started again.
Before starting the vxevac or vxassist move operation, ensure the orphaned subdisk(s) have been reclaimed and iare no longer visible in the vxprint output

# vxprint -qhtg datadg -z
dg datadg   default      default  38000    1524849274.262.fred
 
dm datadg01      purestorage1_0001105a auto 65536 10737352400 -
dm datadg02      purestorage1_00011059 auto 65536 10737352400 -
dm datadg03      purestorage2_000119c7 auto 65536 9017687760 -
dm purestorage4_00011ab7 purestorage4_00011ab7 auto 65536 9017687760 -

v  sasvol -           ENABLED  ACTIVE   29841969648 SELECT -        fsgen
pl sasvol-01 sasvol ENABLED ACTIVE 29841969648 CONCAT -      RW
sd datadg-01   sasvol-01 datadg01 0     10737352400 0      purestorage1_0001105a ENA
sd datadg-02   sasvol-01 datadg02 0     10737352400 10737352400 purestorage1_00011059 ENA
sd datadg-03   sasvol-01 datadg03 0     8367264848 21474704800 purestorage2_000119c7 ENA


CAUTION:
The situation can become even more unstable when attempting the following commands (so please follow the above steps and do not run the following commands listed below):

Impacted disk (purestorage4_00011ab7) associated with TEMP plex "%1"

# vxdmpadm -f disable dmpnodename=purestorage4_00011ab7
 
# vxdg -g  sastempvg -k rmdisk  purestorage4_00011ab7

When attempting to remove the disk from the disk group, VxVM commands can hang or appear hung, pending long timeouts.

# vxdctl mode   (Command hangs)

 

Issue/Introduction

The user is unable to remove temporary plex "%1" for disk group "datadg" for volume "sasvol"
# vxprint -qhtg datadg
dg datadg default default 38000 1524849274.262.fred

dm datadg01 purestorage1_0001105a auto 65536 10737352400 -
dm datadg02 purestorage1_00011059 auto 65536 10737352400 -
dm datadg03 purestorage2_000119c7 auto 65536 9017687760 -
dm purestorage4_00011ab7 purestorage4_00011ab7 auto 65536 9017687760 -

v sasvol - ENABLED ACTIVE 29841969648 SELECT - fsgen
pl %1 sasvol ENABLED TEMPRM 29841969648 CONCAT - WO
sd purestorage4_00011ab7-01 %1 purestorage4_00011ab7 0 8367264848 21474704800 purestorage4_00011ab7 ENA
pl sasvol-01 sasvol ENABLED ACTIVE 29841969648 CONCAT - RW
sd datadg-01 sasvol-01 datadg01 0 10737352400 0 purestorage1_0001105a ENA
sd datadg-02 sasvol-01 datadg02 0 10737352400 10737352400 purestorage1_00011059 ENA
sd datadg-03 sasvol-01 datadg03 0 8367264848 21474704800 purestorage2_000119c7 ENA