- Community Home
- >
- Storage
- >
- Midrange and Enterprise Storage
- >
- StoreVirtual Storage
- >
- Re: Replace an ESXi host & VSA
Categories
Company
Local Language
Forums
Discussions
Forums
- Data Protection and Retention
- Entry Storage Systems
- Legacy
- Midrange and Enterprise Storage
- Storage Networking
- HPE Nimble Storage
Discussions
Discussions
Discussions
Forums
Forums
Discussions
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
- BladeSystem Infrastructure and Application Solutions
- Appliance Servers
- Alpha Servers
- BackOffice Products
- Internet Products
- HPE 9000 and HPE e3000 Servers
- Networking
- Netservers
- Secure OS Software for Linux
- Server Management (Insight Manager 7)
- Windows Server 2003
- Operating System - Tru64 Unix
- ProLiant Deployment and Provisioning
- Linux-Based Community / Regional
- Microsoft System Center Integration
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Community
Resources
Forums
Blogs
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-11-2018 03:46 AM
тАО04-11-2018 03:46 AM
Replace an ESXi host & VSA
Hi Everyone,
I was hoping that I could get some suggestions about the 'best way forward' for a project I'm about to undertake.
The project involves replacing an ESXi host in a 3-host cluster (due to hardware issues). Each of the virtual hosts has a VSA installed, and the 3 VSAs are in a cluster.
The only option that I can see currently is to remove the VSA (on the ESXi host which is to be repalced) from the cluster, leaving the 2 running but not fault tolerant. Replace the ESXi host, and install a new VSA onto that. Then add the new VSA to the cluster, which should then rebuild the array accross the 3 hosts.
Does this sound a reasonable and sensible way forward? I assume the rebuild could take quite a long time.
I have licenses for the 3 VSAs but I'm sure that I read somewhere that the license key is tied to the MAC address of the VSA. Is this the case, and if so can it be worked around just by altering the MAC address on the VSA?
Thanks in advance for any input on this.
Richard
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-11-2018 03:57 AM - edited тАО04-11-2018 03:58 AM
тАО04-11-2018 03:57 AM - edited тАО04-11-2018 03:58 AM
Re: Replace an ESXi host & VSA
Hi,
If you wanna go the difficult and long way, this is indeed the procedure you will need to follow... Disadvantage however is that you will need to rebuild your data twice... And have a longer period of lower availability due to the rebuild times...
Much easier is building the new ESXi host, shut down 1 VSA, move it completely from the old ESXi host to the new one (like a storage vMotion), add the VSA to the inventory by right clicking the VMX file, and start the VSA... If your networking is OK then your VSA will start rebuilding data but only the changes when the VSA was down...
Kr,
Bart
If my post was useful, clik on my KUDOS! "White Star" !
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-11-2018 08:13 AM
тАО04-11-2018 08:13 AM
Re: Replace an ESXi host & VSA
Hi Bart, ok thanks for this information.
This does sound more straight forward and faster than my original option. The networking is slightly different on the new host, as it has a different number of NICs. I should be able to get it configured so that it'll work ok.
Thanks,
Richard
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-12-2018 05:31 PM
тАО04-12-2018 05:31 PM
Re: Replace an ESXi host & VSA
richard,
maybe I missed it, but are you saying the bad node is already offline or is it running but needs to be replaced soon?
Do you have enough storage capacity to run some/all LUNs in NR10+1? I"m going to assume not since it sounds like you ruled out simply gradefully removing the VSA from the cluster (or the VSA is already dead).
The liceneses are tied to the MAC, and at least in hyper-v, you can manually set those and that does work for the licensing.
Also, you don't mention your timing of the removal of one host for the replacement... if you have the replacement online before the other goes off, you can do a node exchange instead and maintain complete redundancy the entire time and not lose any significant downtime (past the time it would take to change the mac address on the VSA after the original one is completely offline.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-13-2018 02:24 AM
тАО04-13-2018 02:24 AM
Re: Replace an ESXi host & VSA
Hi Oikjn,
The node is still online, its just that due to some hardware issue the ESXi host is no longer 'trusted' with a production workload so we want to replace it asap.
We have the replacement hardware connected up and ready to go, but this only has ESXi installed currently and nothing else. The network configurations etc have been set up on this ESXi host, so we can hopefully just swap them over.
This ESXi host isn't currently managed by vCenter due to our licensing restrictions.
The VSA itself is running and healthy with the 3 nodes in NR10. We can run with a VSA node powered off for a while (such as during upgrades) and the LUNs will stay up and accessible.
I haven't ruled out any course of action yet, so I'd be very interested to hear what is involved with doing a node exchange.
Thanks,
Richard
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-13-2018 02:34 AM
тАО04-13-2018 02:34 AM
Re: Replace an ESXi host & VSA
Hi Richard,
Like I told you, since the original server is still online, and so you have access to the VM files, you can shut down that 1 VSA on the server to be replaced.
Copy the directory with all the files from that VSA from the datastore (VMDK, VMX, VSWP and so on) on the old server to the new server. When all files are copied to the new server, you go to the VMX file in the datastore and select 'Add to inventory' when rightclicking that file...
The VSA will be there as a VM on the new server, just check the networking in the VM settings and you are done... Start the VSM, it should be detected by the cluster and all changes of data will be synced automatically...
Kr,
Bart
If my post was useful, clik on my KUDOS! "White Star" !
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-13-2018 07:25 AM
тАО04-13-2018 07:25 AM
Re: Replace an ESXi host & VSA
Bart's option is definitely the quickest to complete and easiest to administer since you don't have to create a new VSA. The only issue is the potential downtime it takes to copy the VSA files from one computer to another. If that is going to take maybe 30 minutes, I can see it being potentially worth the risk, if that is going to take hours, you might not want to risk it, but its all a risk/reward calculation for your and your company. The no-risk way would be to build a new VSA on the new host and then do a "node exchange" to migrate the data of the node to be retired to the new node. This would require just one rebuild to complete, but would also involve zero loss of redundancy to the system.
I can say that for my company, I"d probably go with Bart's approach because our VSAs are small and the cost of a loss would be minor, but that might not work for everyone.
My nearest analogy situation might be where you were standing on one side of a 2-lane road and needed to get to a store on the other side, would you just look both ways and run across the road when it appears clear or would you walk to the closest corner? What if it was a 6-lane divided highway?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-13-2018 12:11 PM - edited тАО04-14-2018 02:56 AM
тАО04-13-2018 12:11 PM - edited тАО04-14-2018 02:56 AM
Re: Replace an ESXi host & VSA
you mean the risk of a second VSA failing during the copy process?
I think the copy will take about 10 hours.
*edit*
Ok I was wrong about the time it will take to copy. It is about 50% done after 12 hours. I'm using a Veeam copy job to copy from one host to the other, I tried to use the 10gb interfaces but either it isn't do that or there's another bottleneck somewhere. It's copying 8TB of data, roughly 105MB/sec.
Richard
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО04-16-2018 11:52 AM
тАО04-16-2018 11:52 AM
Re: Replace an ESXi host & VSA
105MB sounds like a 1GB link or something that is getting restricted at that speed.
yes I was referring to the unprotected state during that copy time period + whatever rebuild is required afterwards. Sounds like you are comfortable with that day downtime so the entire VM is certainly the way to go.
Are you using Veeam to backup the VSA nodes or just using VEEAM for this copy job? Why didn't you use the replication side of Veeam to replicate it over to the new host with even less downtime?
We use Veeam too (but for Hyper-v). I"m pretty sure they suggest NOT backing up the VSA itself since there really is almost no use case to justify it. You could probably cheat some setup recovery time by backing up the VSAs while excluding the main stroage vmdk, but as you can see, the time to restore the full VSA raw storage is typically too long to be practical. Each individual VSA is meant to be disposable so there is no "need" to back them up since if one dies you simply make a new one and do a node exchange to replace the bad node with the new one.