Not maintained in Debian

[bzed]

collectd is actually not maintained in Debian at all.

I intend to adopt the package if nobody wants to take care of it anymore.

[zebity]

Hi @bzed , not sure what you have in mind here.

I started work on collectd as debian based Ubuntu had defect (collectd/collectd#3143), as a result of QEMU moving forward and collectd staying locked at 5.8.1 .

As part of fixing this we ended up having to go though large portion of collectd code to allow it to compile with new Ubuntu tool chain (collectd/collectd#3153).

All this took some time to flow into master and I was hoping that once a 5.9.x release was cut (this is being looked after mostly by @mrunge who has been incredibly helpful in getting all this code into order) then could get all these fixes and updates flowing into Debian and from there get picked up by Ubuntu build process. We have hit a few bumps along the way due to some defects and additional features that have continued to flow into 5.9.x target.
But things are close ...

I believe that all the collectd coding work should be done via collectd/collectd and then Debian build should just grab the appropriate sub-set of plugins + add required systemd or other startup scripts (though these could also come via upstream) and package these up into APT bundles.

Now that we have fixed the tool chain issue, there is no need to change the build parameters to build Debian / Ubuntu packages.

So to your point about Debian not being maintained, what work is required ?

Very happy to help push this Debian / Ubuntu release work forward as I need a 5.9.x package cut sooner rather than later (as build and installing collectd from source on target machines is not viable).

Also fyi, as part of working on code I created some helper scripts and tools which I have documented here: https://tips.graphica.com.au/tool-and-tips-on-building-collectd/

Lets get some momentum going on this ...

Cheers

Zebity.

[zebity]

Hi @mfournier and @tokkee ,

could you please chime in on this.

Some guidance on getting Debian updated to 5.9.x would be helpful.

The collectd repository has just had 5.10 tag added and there are a lot of updates that have gone into 5.9.x that need into flow in Debian / Ubuntu build processes and releases.

Cheers,

Zebity.

[zebity]

@bzed have taken affirmative action and submitted pull request for 5.9.1 candidate (#18).

I say candidate as am new to creating debian package and have essentially hacked out update based on reading docs and testing.

A more experience set of eyes over this would be helpful.

Cheers,

Zebity

[bzed]

@zebity the problem is that neither @mfournier nor @tokkee react on mails/irc or on the open RC bugs in Debian. Actually I'm planning to upload 5.9.1 to Debian as soon as I have some spare time for it. The current state is a mess, thats why I've opened this bug...

[mrunge]

I would avoid collectd 5.9.1, there are multiple reports of broken plugins.

5.9.2 basically needs tagging, the changelog is pretty much ready, and after the tag, the tarball should be uploaded. That's basically what's left for 5.9.2. If you could give the 5.9. branch a try, that would speed up 5.9.2 release.

[zebity]

Hi @bzed ,

I included you in discussion on release process: collectd/collectd#3223

As I believe that there is a core set of people that are interested in getting more frequent and responsive process going with collectd.

I have also been looking at Debian maintainer model and adoption process.
If we cannot get past 8.x which is now a relic then I don’t see any other option.
I would be happy to join Debian as part of collectd maintenance process, but the process appears onerous and requires sponsor.
As an existing Debian member, you would have to start the ball rolling.

I am sending this to you via pkg_debian issue you’re raised as this consistent with your original issue, and it is best to be open and frank.
I have no personal contact with current maintainers @mfournier & @tokkee and no knowledge of their current circumstances.
They could either be very occupied with other things or simply worn out from labours keeping the release cycles going.

I believe a key objective should be to further automate the release process to reduce the burden.

Cheers,

Zebity.