Skip to main content
Pure1 Support Portal

ActiveCluster: How To Replace A Permanently Lost Or Destroyed FlashArray

Problem

Following the permanent loss or destruction of one of the two FlashArrays supporting stretched pods there are a few cleanup steps that will need to be run on the surviving FlashArray. 

Note: This procedure only applies to pods with an online status on the winning FlashArray (mediation winner) for each stretched pod. The procedure will not work for pods that show offline at the surviving FlashArray.  For pods that have an offline status on the surviving FlashArray, please work with Pure Support and follow the standard pod clone procedure for offline pod data (as needed).   

Impact

While the destroyed FlashArray shows a 'Connected' status on the surviving FlashArray it will not be possible to establish a second synchronous connection to a (new) replacement FA in order to rebuild ActiveCluster.

Solution

The following steps must be performed:

Step 1. Determine the status and remote host connectivity for each pod on the surviving array.

From the CLI:

Use purevol list --remote --connect 

pureuser@tmefa04> purevol list --remote --connect  
Name        Size  LUN  Host Group  Host            
Pod1::Vol1  444G  1    -           tmeesx40        
Pod1::Vol1  444G  1    -           tmeesx50        
Vol2        777G  2    -           tmeesx40        
Vol2        777G  2    -           tmeesx50        
Pod1::Vol1  444G  1    -           tmefa05:tmeesx40
Pod1::Vol1  444G  1    -           tmefa05:tmeesx50

From the FlashArray GUI:

Navigate to Storage > Volumes > <Volume Name>.  From the options button at the top right of the Connected Hosts pane, select Show Remote Connections.

 Screen Shot 2019-08-27 at 7.06.26 PM.png

 

Step 2. Disconnect all remote host and/or remote host group connections to the stretched pod volumes.

From the CLI:

Use purevol disconnect <destroyed-array-name>:<remote-hostname> <pod-name>::<vol-name>

pureuser@tmefa04> purevol disconnect --host tmefa05:tmeesx40 Pod1::Vol1
Name        Host Group  Host            
Pod1::Vol1  -           tmefa05:tmeesx40
pureuser@tmefa04> purevol disconnect --host tmefa05:tmeesx50 Pod1::Vol1
Name        Host Group  Host            
Pod1::Vol1  -           tmefa05:tmeesx50

From the GUI:

From the Connected Hosts pane that was opened in Step 1, click on the x at the right of each remote host or remote host group connection.  Repeat for all remote host / remote host group connections to volumes within the pod.

Screen Shot 2019-08-27 at 7.06.53 PM.png


Step 3. Unstretch the pods from the destroyed FlashArray back to the surviving FlashArray.

From the CLI:

Use purepod remove --array <destroyed-array-name> <pod name>

pureuser@tmefa04> purepod remove --array tmefa05 Pod1            
Name  Source  Array    Status  Frozen At
Pod1  -       tmefa04  online  -    

From the GUI:

Navigate to Storage > Pods > < Pod Name>.  From the Arrays pane, click on the Trash Can icon next to the name of the destroyed FlashArray and then confirm the removal of the array.  Repeat for all stretched pods. 

Screen Shot 2019-08-27 at 7.47.48 PM.png
 

Step 4. Disconnect destroyed FA

From the CLI:

 Use purearray disconnect --local-only <destroyed-FA-name>  to complete the removal.  This command can only be run by Pure Support.   

root@tmefa04-ct0:~# purearray disconnect --local-only tmefa05
Name
tmefa05