Disabling the pdcache Cache of Specified Nodes in Batches

Hi team, here's a new case.

Problem Symptom

During the inspection, the check result of item Checking Whether pdcache Is Enabled on the SAS3508 RAID Controller Card of the Node for multiple nodes is Not passed.

Problem Diagnosis

If pdcache is enabled and the check result of this item is Not passed, you can disable pdcache on the nodes as instructed in this section.

Causes

If a node uses the 3508 RAID controller card and pdcache is enabled, services on the node may be affected.

Solution

1. Use SSH to log in to the active FSM node as user root and run the following command to go to the path for operating pdcache on multiple nodes:

   cd /opt/dsware/manager/script/pdcache/

2. Run the vim storage_hosts command to edit storage_hosts in the directory. Enter the management IP addresses of the nodes where pdcache is to be disabled in the format shown in the following figure, save the settings, and exit.

   

   
   

   The prerequisite for batch change command is that the passwords of the target nodes are the same. If the passwords of some nodes are different, classify nodes based on different passwords, and change the passwords of the nodes in batches.

3. Run the following command to disable the pdcache cache on specified nodes in batches:

   The following is an example of the command for disabling the pdcache cache on multiple nodes in batches:

   

   1. sh execBatch.sh;

   2. When please input user: is prompted in the command output, enter the username for logging in to the target node using SSH.

   3. When please input user's password: is prompted in the command output, enter the password of the user for logging in to the target node using SSH.

   4. When please input root's password: is prompted in the command output, enter the password of user root for logging in to the target node.

   5. When please input your action: is prompted in the command output, enter handle.

4. Check switch off pdcache success_host at the end of the command output. The IP addresses of the nodes where pdcache is successfully disabled are displayed.

   If information similar to the following is displayed, pdcache on the nodes is successfully disabled in batches.

   

   Note: If some IP addresses under the switch off pdcache fail_host field in the command output fail to be executed, possible causes are as follows:

   1. Log in to the failed node using SSH.

   2. Run the /opt/MegaRAID/storcli/storcli64 /c0/v0 set pdcache=off command to disable pdcache.

   3. If Status = Success is displayed in the command output, the execution is successful.

   4. The ssh user name or password of these nodes may be incorrect.

      Fill the failed nodes in the storage_hosts file again and find the correct user names and passwords for logging in to these nodes using SSH.

      Run the sh execBatch.sh command again.

   5. If a few failed nodes still exist after you run the command again, perform the following operations:

Check After Recovery

After the fault is rectified, you can use SmartKit to perform an inspection again.