event plugins update?

Deadline
1

I want to see what the logs would look like if my Python code contains error so I removed a few quotation marks from my script. It should had caused errors, however the log still printed out normally for quite a while before it actually stopped working.

Is there a way to force them to pick up the new code right away?

2

Event plugins sync every 5 mins:
docs.thinkboxsoftware.com/produ … ervals.png

You can lower the 300 seconds value, I just wouldn’t go too low, as it could negatively impact your farm.

3

You can force the event plugin refresh in the monitor by going to Tools->Synchronise Scripts & Plugins.

4

Is there conditions for the sync to happen? If the slave is busy will the sync be done directly after?

I’m currently using Deadline9. This morning I updated an event and looked at the jobs in the afternoon, some slaves were up to date others were not.

As for the Tools->Synchronise Scripts & Plugins. Does it updates the scripts and plugins for all of the slaves?

5

“Tools->Synchronise Scripts & Plugins” is only for the Monitor’s own scripts and plugins.

The plugin reload process changed a bit. A little overview is here:

There is a duplicate thread here with more detail however:

Considering you’re the second person who has raised the issue, I’ll go back and update the dev ticket about it.

7

Do we have access to the dev ticket?
What is the ETA?
Do we have a way to force the slave to reload the event plugins? Restart the slave, restart the machine…?
Is there a place where we can see the last time it updated the events?

Sorry for the multiple questions :wink:

8

As I understand it, restarting the slave will update all event plugins and scripts on it.

1 Like
9

Yeah, restarting the Slave (easy if you have remote administration enabled) is the way to go here.

The developer tickets aren’t publicly available but you can always ask us to poke on them.

If you’re developing scripts and need to restart often, running a Slave on your workstation or a test machine is the best way to iterate quickly on changes. Additionally, having a local Repository you can develop on as a staging area is quite nice and what we all do over here. @mois.moshev has had some good luck using this workflow.

1 Like
10

Would be nice to have a log somewhere when events are reloaded.

Thanks a lot for your help guys.

11

It usually says in the applications log if verbose logging is enabled. I went digging to find an example, but came up empty handed. It’s something like “reloading X because it has changed on disk” where X is an event.

12

The log message saying that an event plugin is being reloaded as its changed on disk appears in your “Console” panel in DL monitor if you have Verbose Logging enabled in Repo Options.

Developers/Scripters in DL should always have the “Console” & “Background Ops” panels docked in their Monitor UI as power tools. :wink: HINT: you could saved a monitor layout either locally or share globally in your pipeline to ‘switch’ into dev mode.

1 Like
13

I think it should be noted in the docs: Event Plugins — Deadline 10.1.18.5 documentation

I’ve been confusing for a while till I found out this thread.

14

I too would be interested in more details about this.
Currently I am setting up a Linux worker, from my Windows machine. The local Windows worker always reloads its plugins when restarted, but this doesn’t seem to happen reliably on Linux.
Perhaps this changed with sandboxing? From what I gather, event plugins run in their own sandbox process, and it is unclear how to manage that.