Storage Replication & High-Availability

Aug 30, 2018
3
0
1
31
Hey,

Im testing the features storage replication and ha but I don't quiet understand this sentence in the wiki:

https://pve.proxmox.com/wiki/Storage_Replication

High-Availability is allowed in combination with storage replication, but it has the following implications:
  • redistributing services after a more preferred node comes online will lead to errors.
  • recovery works, but there may be some data loss between the last synced time and the time a node failed.
I tried the following:
3 node cluster, local zfs storage
lxc 18.04 container with webserver on node1, replication to node2, every minute

Test1) Manual migration from webserver on node1 -> node2, no problems, replication task changes from node2 -> node1 as it says in the wiki

Test2) webserver on node2, replication to node1 every minute, force shutdown of node2, manual migration like "Migrating a guest in Case of Error" -> works perfectly, after the migration replication task changes from node2 -> node1 to node1 -> node2

Test3) Enable HA, webserver on node1, replication task to node2 every minute, force shutdown of node1, after 150 seconds webserver gets started on node2, replication task fails because node1 is offline (as ecpected). Start of node1 (after "repair"), webserver disappears on node1 (I guess because it sees that it is running on node2), offline migration of webserver from node2 -> node1, works without problems, replication task changes from node2 -> node1 to node1 -> node2

So where exactly does redistributing services after more preferred node comes online lead to errors?
 

Attachments

  • upload_2019-5-21_14-16-36.png
    upload_2019-5-21_14-16-36.png
    2.9 KB · Views: 9
https://pve.proxmox.com/wiki/Storage_Replication

High-Availability is allowed in combination with storage replication, but it has the following implications:
  • redistributing services after a more preferred node comes online will lead to errors.
  • recovery works, but there may be some data loss between the last synced time and the time a node failed.

That means: In case of a VM using "replicated" storage HA works in principle in the same way as in case of a "shared" storage, but live migration is not possible. But when tries HA live-migration? Answer: When - after caused by failing of all nodes in the assigned HA group - a HA VM runs in a node which is not its HA group and one of the nodes in the HA group comes back HA tries to migrate to that node and this fails (because live migration not possible). That's not an error in terms of something is destroyed, but just attempting (every 10 seconds) a feature which does not exist.
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!