Zammad tmp size

Technical assistance
1

Infos:

  • Used Zammad version: 4.0
  • Used Zammad installation type: package
  • Operating system: Debian

Stats:

  • No of agents: 108
  • No of users: 79054
  • No of overviews: 43
  • No of groups: 36
  • No of new tickets per day: ~1700

I am considering moving /opt/zammad/tmp to RAMDISK. I would like to ask you guys what is the size of your tmp directory. Mine is 20G which quite big I think.

The biggest top folder in tmp is cache_file_store_production ~18G

Do you know why it is so big? Should I reserve at least 20G of RAM for the RAMDISK?

1 Like
2

While you can do so please keep in mind that Zammad stores unprocessible mails (if applicable) within tmp directory too. That means: If your host reboots for whatever reason and you didn’t ensure the directory was empty, the files would be lost. That’s technically the nature of the ramdisk.

The size of the tmp directory, especially the cache directories within it, highly depend on the usage of your Zammad instance. The numbers you provided are fairly big so Zammad caches a lot more than it would on other instances. Removing the cache directories is perfectly safe in case you have to!

I’d recommend to do that during upgrades at least.

3

In such case, i think its better to user ssd mirror for tmp.

4

@konrad I’m curious to know what your Zammad server specs are for running at that level of usage.

We only have about 40 agents, maybe 2000 users, 25 overviews, 25-100 tickets/day.
But we seem to have performance issues regularly.

1 Like
5

Our setup is split between 3 machines.
We have:

  • Zammad 1 (16 cores, 12 GB RAM, 75 GB SSD)
  • Zammad 2 (16 cores, 8 GB RAM, 50 GB SSD
  • Postgresql (4 cores, 16 GB RAM, 400 GB SSD)
  • Elastic (4 cores, 12 GB RAM, 50 GB SSD)

We have two Zammad instances used by separate entities. “Zammad 1” is the machine that handles that traffic:

  • No of agents: 108
  • No of users: 79054
  • No of overviews: 43
  • No of groups: 36
  • No of new tickets per day: ~1700

We have common machines for Database and Elastic (for both Zammad instances).

Did you play with that parameters?

  • WEB_CONCURRENCY
  • ZAMMAD_SESSION_JOBS_CONCURRENT

I had to adjust them and everything started working much better.

6

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