Import Archive Timeout

[Rob Byrne]

Hi Everyone,

We recently began a migration from OracleDB to Postgres in our environment with our Rundeck VM. I exported the project archives for both of the projects we had, and am now trying to import them in on the postgres DB. This is where I'm running into issues.

The first project imported using the "Import Archive" after a minute or so and went fine (beyond some jobs missing execution history, but that's a different issue). However, the larger of the two projects is failing to import properly. Every time I've attempted to import via the Web UI; it times out after about a minute or so (Gateway 504 Timeout).

We are using Nginx as a web proxy, so I've tried adjusting our Nginx config to allow for a longer timeout. Despite changing this, it continually times out at about the same point each time. Is there an internal timeout for Rundeck I need to adjust?

I've also tried using the CLI, but it's indicating our API version is out of date. (Is this controlled by something other than just the rundeck version?)

Any help would be appreciated. Thanks!

[Rob Byrne]

After letting it run for 4 hours on an import (via RD-CLI) it failed out with connection timeout again, but now the jobs are listed in the web UI (and they execute normally). The only issue is I can't seem to modify these jobs without getting a long loading time and a connection timeout (504 again). Is it possible there's some issue with the UUID these jobs have? If so, can that be re-generated somehow?

[JP Lassnibatt]

Hey Rob, those exported projects have the job's logs too? Any chance you can export/import without previous logs (skip the "Executions" box)?
I think it's not related to UUIDs, but to regenerate, you need to select "Remove UUIDs" in the import process.

[Rob Byrne]

Hello,

I checked in on the server, today, and disabled the SCM import plugin we weren't using. Either that, or letting it sit long enough (perhaps the Postgres DB had some behind-the-scenes work it was doing?) Seemed to fix the issue. We are now able modify the jobs without issue, and without needing to re-do the import.

I'll keep an eye on it in case other issues crop up, but it seems to be functioning as expected right now.

Thanks for the troubleshooting advice! That was my next step if it hadn't worked this morning.

Rob