Is it possible to configure a failover cluster for the EasyMorph Server?

Server
1

Dear,

In our organization we are using EasyMorph within more and more projects / solutions… so we are depending more and more upon the EasyMorph Server. However we did not have any issues regarding its availability, we have some concerns about this single point of failure.

So I have searched for any references regarding setting up a failover cluster or something similar, but I couldn’t find anything.

Does anyone have some information regarding this?

Kr
Edward

2

High availability and clustering are definitely on the roadmap for EasyMorph Server, but are not available yet out of the box.

There are no ideal solutions for high availability at this point. There are various ways how to increase the availability of EasyMorph Server, but none of them is perfect. A few ideas:

  • A cold backup copy of the Server based on a backed up image of the machine’s hard drive
  • A hot backup copy of the Server with shared Server system folder (this variant is highly sensitive to latency of the shared drive)
  • Have two (or more) separate Server installations, split spaces between them (if one goes down, at least the other keeps working, so the damage isn’t 100%)
1 Like
3

Hi @dgudkov,

Thank you for your fast reply. Nice to hear that this is on the roadmap. Can you give an estimate (roughly) on when we can expect something? So I can give feedback within my organization regarding this and if needed asses if we have to take a look at (or build) a workaround.

Kr.
Edward

4

We’re currently working on adding users, user groups, and roles, as well as support for external identity providers. This is expected to go into production in July this year. After that, we will introduce task triggers, probably in October. At the end of this year, we will start drafting the technical architecture of Server clustering. The clustering itself might become available somewhere around the summer of next year.

1 Like
5

In my big company we have one active and one passive server with load balancer in front of them. We would love the clustering feature.

1 Like
6

How do you sync projects and data files across two servers?

Do you use some fault-tolerant distributed object storage like Ceph or a fault-tolerant file system for projects and data files?

7

For data/projects, we use NAS. For system files we were using NAS but it appears to be problematic with a lot of users/tasks => we are reverting to local folder and we will just script copy of files (task config mainly)

8

My understanding is that after the workload was switched from Server A to Server B, then script copying should run on Server B.

I’m curious if you do regular (e.g. monthly or quarterly) “rehearsals” - i.e. switch from one Server to another Server to make sure the setup works if it becomes necessary?

9

Copies are done frequently, not only when there’s a switch

Generally we do one test per year, I don’t remember the results of last year test but I remember there were 1 or 2 necessary fixes. But I must say that it’s globally easy to switch, using NAS or different local folders that you synchronize.

1 Like