Ram usage in rake:rebuild

  • Used Zammad version: 4.1
  • Used Zammad installation type: source
  • Operating system: Ubuntu 18.04
  • Browser + version: Chrome

Expected behavior:

zammad run rake searchindex:rebuild
should rebuild the elasticsearch index.

Actual behavior:

It goes through most, but when it hits tickets it gets maybe 1/3 to half of the way through before using up all the ram on the system. I have 10gb allocated to the server, and it never goes above half of that under normal loads.

Is there anything I can do to alleviate this ram use creep when rebuilding?

Steps to reproduce the behavior:

zammad run rake searchindex:rebuild

You’re looking at your hypervisors ram usage of the VM in question there.
This is -by far- the least helping graph you can look at.

Because your OS caches files it takes in your memory as long as possible.
This fills your memory technically and thus looks like high memory usage for your hypervisor.

Instead of doing above, run an actual top or htop on the machine.
You’ll be suprised what actually goes on.

I can’t explain it, but I’m glad anyways.
It doesn’t crash out if I have htop open in another putty session.
Real mem usage is under 2 gigs, never realized how poorly that hypervisor graph represented it.


This topic was automatically closed 120 days after the last reply. New replies are no longer allowed.