Solved

Limit on concurrent migrations?

  • 27 April 2020
  • 2 replies
  • 3174 views

Userlevel 1
Badge +1

My migration team is doing a good job trying to push as many VM’s through Move to stage reboots after-hours.  I’ve noticed that replication slows way down when we have about 15-20 VM’s waiting for cutover. 

I have checked the load on the Move VM and increased the amount of CPUs.  Replication goes well (25Gbit network), but now we only have 2 seed streams going with 19 waiting to cutover.  I have 4 other VM’s that are in the “seeding data” phase, but haven’t started their copies.  Is this a bottleneck within the Move postgres database?

Any ideas on how I can get more VM’s copying in parallel?

Thanks,

icon

Best answer by David Teague 23 June 2020, 21:08

View original

This topic has been closed for comments

2 replies

Badge

Sorry for the late reply - From personal experience, the number of concurrent migrations that happen at the same time seems to be tied to the amount of data you are moving. I was doing 10 migrations at a time when I was a customer (at my job before nutanix) and it was around 1-3 TB of data without a problem. If I have any large VM’s 10+ tbs on their own I would do that migration on its own without any other migrations running. I will ask the Move team if there is a guideline and get back to you.

Badge

@BlakeRobertsNBL  Here is what the move team had to say about limitations.

 

Move performs up to 16 disks migration in parallel. If it is from the single ESX host, then the limit is 8 disks in parallel.

 

Hope this helps.