Correct way to uninstall a content pack

Suspect that something may be amiss. Google’d, checked docs, etc. For example, take any of the ‘out of the box’ content packs - install one. They create (amongst other things) table, cache adapters and data adapters. On the flip side, if you try to delete/uninstall the content pack - it appears that something doesn’t happen correctly - because the server logs will fill, fast and hard with a metric ton of warning messages. Even more interesting is that any lookup table that [apparently] was created after that content pack was installed and removed prior to trying to remove the content pack - is now “stuck”. eg: you remove that lookup table (no log issues/errors) but when you try to remove the content pack, that other lookup table now adds even more fury to the logging action occurring with the server log.

Noticed another interesting side effect - one can ?“double install”? the same content pack… Would think that you could only have one “version” of a content pack installed.

Have a suspicion that when a lookup table/cache adapter/data adapter is removed from graylog - that it’s still “stuck” inside mongodb (at least from what can be gleaned of how things work). If that’s correct, how does one clean that up? How is this process expected to work? Unfortunately, there’s simply not enough information in the graylog documentation. Hoping someone’s been down this path previously and has some good pointers on how to correct - short of uninstall and re-install.

Thanks!

I will not help you with cleaning but for my experiences, content packs and illuminate packs install/deinstall is buggy.

They often leave garbage who are not visible in the UI.

I think you will need to dig the MongoDB database to clean up.

Completely understand on not wanting to get drawn into the cleaning part. But I will happily accept anything you might suggest one be aware of. :wink:

I’m somewhat surprised at just how much ‘garbage’ is left behind and how “uninstallable” things are. Was even considering “wiping” the entire install. Extractors can be easily exported. Pipeline rules aren’t “too bad” to capture (especially when their hand written/not crafted with the wizard interface). However, the multitude of dashboards and the reports therein - there doesn’t appear to be an easy means to “export” all the tabs + associated reports per tab, to easily re-import.

On one hand - really, really like what Graylog can do. On the other, it’s irritating to get it “right” and then keep it right. Especially if you want/need to explore and try different scenarios that -should otherwise- be possible without a complete rebuild to “undo”.

While familiar with some SQL engines - not familiar with MongoDB. Any tool/s that you found valuable when trying to review what was in MongoDB?

Thanks!

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.