Note that the requirement in the Site Prep Guide is against using ANY XtremIO storage for hosting the virtual XMS (and not just the one that the XMS is managing).
We have had some customer cases in the past where the shutdown/startup activities on an XtremIO array was made complex or got delayed because the customer had chosen to store the vXMS on an XtremIO array.
If you ever find yourself in that situation like I have where the vXMS was stored on the XtremIO and it was shutdown, you can use VMware player locally to deploy another vXMS, install the code your array is at, and perform XMS recovery. Once that is done, you can issue the start-cluster command.
I was working with a customer whose only storage was the XtremIO. I moved the vXMS off to a local host storage prior to our planned activity and they had moved back to the XtremIO without me noticing. That's how I ended up in the scenario mentioned and what I did to recover.
Sent from my mobile device. Please excuse any mistyped or abbreviated words.
Chris Mahon, VCP, EMCCE, EMCIE
EMC Corporation | Delivery Specialist II
We are running XMS 4.2.1. We currently have 2 x XMS in different countries. We have 2 x XtremIO XBricks running 4.0.15-24.
We deployed 2 x XMS which managed each country seperately. Key reason the 4,096 volume XMS limit.
We now need to do a data centre powerdown in one of our locations. This presents some dependancy challenges as XMS is a VM. The XMS is stored on VNX8000 but we have a dependancy now on the ESX estate remaining up while we powerdown XtremIO from the XMS.
I have the following questions which I hope someone on this thread can answer.
Is it possible to safely shutdown XtremIO without XMS? (not emergency shutdown)
Any updates on XMS and 4096 volume limit?
Is it possible to temporarily move management from 1 XMS to another? Would this introduce 4,096 volume limite?
Thanks very much for any information.
I am replying to my own post just in case it helps anyone else who has similar problems. I had email conversations with Chris Mahon offline. Below is summary of these conversations and my own research / practice.
It is not possible to shutdown XtremIO Cluster cleanly without XMS.
In our case we can remove dependancy on XMS in one location by change the XMS that manages XtremIO to our remote XMS. This is a temporary arrangement so powerdown can be done. We have RecoverPoint also so the array settings must get updated in RecoverPoint.
Procedure is below with IP addresses and names obscured.
Remove Location 1 XtremIO from Location 1 XMS
Logon as xmsadmin
Change to admin
Remove cluster from XMS
xmcli (admin)> remove-cluster cluster-id="**-XTREMIO-****"
Are you sure you want to remove cluster **-XTREMIO-*** ? (Yes/No): Yes
Cluster **-XTREMIO-***  removed successfully (Wait 5 minutes here)
Add Location 1 XtremIO to Location 2 XMS
Logon as xmsadmin
Change to admin
Add Location 1 XtremIO cluster by using one of the Storage Controller IP address
xmcli (admin)> add-cluster sc-mgr-host="XX.XXX.X.XX"
17:07:09 - Collecting cluster information
17:07:19 - Loading data from cluster
Clusters added successfully (wait 5 minutes here)
Log onto Irish XMS Gui and the Location 1 XtremIO Cluster is available in drop down as well as Irish XtremIO.
Follow similar procedure to move Location 1 XtremIO back to Location 1 XMS.
Following this procedure doesn’t inforce the 4,096 volume limit. This will be in place only while XMS manages both 2s XtremIO Clusters.
In RecoverPoint Change RecoverPoint XMS Server to new XMS Server.
In GUI go to RPA Clusters, Location 1 Cluster, Storage
In XtremIO entry do the following Edit
Change logon credentials to Location 2 XMS (or Location 1 if moving back). Supply new XMS IP address and credentials.
Hope that helps,