Friday, October 4, 2013

PowerHA 7.1.x Migration • ERROR: PowerHA System Mirror requires a shared disk connected to all nodes in the cluster. None could be found.

Environment

AIX 6.1 TL 08 SP 00
+ POWERHA 6.1 SP 09 and SP 11

AND

AIX 6.1 TL 08 SP 02
+ POWERHA 6.1 SP 09 and SP 11

Problem Description

When we try run ‘clmigcheck’ script to perform validation and verification of the cluster for migration of POWERHA to newer version ( PowerHA 7.1.X), the script fails when we try discover any available shared disk to create new repository disk on both nodes.

For more information about upgrade process please read this link below:

Upgrading an offline cluster from PowerHA SystemMirror 6.1, or earlier, to PowerHA SystemMirror 7.1.0, or later


Action Taken

We ran the ‘clmigcheck’ script until occurred the error:

# /usr/sbin/clmigcheck
——————[ PowerHA System Mirror Migration Check ]——————-

Please select one of the following options:

1       = Check ODM configuration.

2       = Check snapshot configuration.

3       = Enter reopsitory disk and multicast IP addresses.


Select one of the above,”x”to exit or “h” for help:  1
—————————————————————————————————
[…]

——————[ PowerHA System Mirror Migration Check ]——————-

The ODM has no unsupported elements.

Hit <Enter> to continue 
—————————————————————————————————
[…]

——————[ PowerHA System Mirror Migration Check ]——————-

Please select one of the following options:

1       = Check ODM configuration.

2       = Check snapshot configuration.

3       = Enter reopsitory disk and multicast IP addresses.


Select one of the above,”x”to exit or “h” for help:  3
Verifying clcomd communication, please be patient.


ERROR:  PowerHA System Mirror requires a shared disk connected to all nodes
in the cluster. None could be found.
  • If you check the log this execution you probably will see the same exit described here:

# cat /tmp/clmigcheck/clmigcheck.log
——————————————————————————————————-
PowerHA System Mirror migration check program.
Run date: Thu Jul 18 10:29:49 BRT 2013 on machine ossaix61a

[…]

ck_sites: node_list is: ossaix61a ossaix61b

cllssite: No instances found in the HACMPsite class
cllssite: Site display operation was not successful.


ck_sites: Cluster node list = ossaix61a ossaix61b
ck_sites: Local site =
ck_sites: Local site nodes =
ck_sites: Remote site =
ck_sites: Remote site nodes =
ck_sites: First node at remote site =
prompt_disks: Called

prompt_disk: no sites defined, prompt for one hdisk

prompt_disks: Output from cl_lspvids -n ossaix61a ossaix61b:

# No free disks found.
# Check each node to see if any disks need to have PVIDs allocated.


prompt_disks: List of local PVIDs is: 00f6ba989c25c3e9|00f6ba989c25c420|00f6ba989c45d6e1|00f6ba989c45de8e|00f6ba98f1d3da0f|00f6ba989c45e986|00f6ba981ef6c45c|00f6ba981ef6c502|00f6ba981ef6c59c|00f6ba981ef6c63d|00f7f921ad6672ee|00c4f685c3f6b553|00f7f9213df83f24|none|none|none

prompt_disks: List of shared PVIDs is: 

prompt_disks: uniq list of shared PVIDs is: 

prompt_disks: Shared disks connected to each node in the cluster is:


ERROR: PowerHA System Mirror requires a shared disk connected to all nodes
in the cluster. None could be found.

Error termination on: Thu Jul 18 10:29:56 BRT 2013
——————————————————————————————————-

  • Double check the nodes using any command (lspv, lquerypv or readvgda) to verify if the same shared disk exists on all nodes. If there is a shared disk over the nodes but the ‘clmigcheck’ script doesn’t find it, you can change it to correct this problem. Search for the line exposed below on  ‘clmigcheck’ script and perform the change like described:

FROM:

# remove duplicate PVIDs
pvid_list=`echo ${raw_list} |tr ’ ’ ‘\n’ | sort | uniq | tr ‘\n’ ’ ‘`

TO:

# remove duplicate PVIDs
#pvid_list=`echo ${raw_list} |tr ’ ’ ‘\n’ | sort | uniq | tr ‘\n’ ’ ‘`
pvid_list=$(/usr/es/sbin/cluster/sbin/cl_lspvids | egrep -e “`lspv | tr -s ’ ’ | cut -f2 -d’ ’ | paste -d ‘|’ -s -`|all selected nodes|on node ${local_node}” | cut -f3 -d ‘:’ |tr ’ ’ ‘\n’ | sort | uniq | tr ‘\n’ ’ ‘)

  • After this you have to execute the script ‘clmigcheck’ again, but now you will be able to find all available shared disk candidates to create the repository disk and move on with the upgrade.

# /usr/sbin/clmigcheck

——————[ PowerHA System Mirror Migration Check ]——————-

Please select one of the following options:

1       = Check ODM configuration.

2       = Check snapshot configuration.

3       = Enter reopsitory disk and multicast IP addresses.


Select one of the above,”x”to exit or “h” for help:  3
Verifying clcomd communication, please be patient.
—————————————————————————————————

[…]

——————[ PowerHA System Mirror Migration Check ]——————-

Select the disk to use for the repository

        1       = 00f6ba989c45e986(hdisk5)
        2       = 00f6ba98f1d3da0f(hdisk4)

Select one of the above or “h” for help or “x” to exit:  
—————————————————————————————————