My x86 Gentoo buildbot worker is down until further notice due to a dead power supply fan. Given the age of this machine, I'm not sure that it's even worthwhile to replace the power supply, though I do plan to bring the worker back in some form as resources allow.
Unfortunately, this worker provided three unique functions to the buildbot fleet, namely 32-bit (x86) builds, installation tests, and refleak checks. Does anyone else have the capacity to pick up some of the slack until this worker is back, on either new or existing worker(s)?
Last summer I started my buildbot for a couple or reasons:
a) to have it happen and work on a server I managed so I could examine why things were failing b) show commitment to watching the behavior of a bot
For the record, I have commented - and the "bot management" side of things has responded quickly when something that is not the fault of the bots themselves occurs - and gets posted. My thanks for that.
When I posted that I was going to start a bot a number of core-developers commented, on the side, that while it was "nice" to be starting a new bot, and "nice" that I would be paying attention what was really needed was patches to fix the failing tests.
For two+ months I invested nearly every free moment I had (and sometimes vacation moments when I should not have) I worked, and resolved all the failing tests for the platform of my bot - both those that were platform-centric as well as compiler centric.
I understand that core developers are busy people -also investing lots of their free time - and my platform is not on their interest list. But this becomes a self-fulling prophecy - no interest becomes the platform fails - and the platform continues to fail because proposed patches either never get reviewed, never get commited after review, or a second reviewer comes along and wants it done differently, and then it still lies there.
I am not intending to be critical of anyone. I am grateful for all the time and assistance I have received. However, if the combined efforts never result in getting all the platform test failures repaired there is no pint in continuing the bot.
Besides time there is also the fiscal cost. I recently computed that the rising cost of electricity has brought the expense of keeping the server powered-on 24x7 more than I care to bear, and once I have a cheaper replacement for the services I host (a PC or even a raspbery-pi) I will be powering the server off, and most of the time the bot will be down.
Regardless of the status of the bot - I'll continue to monitor the PR's I have submitted and even fireup the bot again, should there be a reason to. But my time is extremely limited now (end of quarter projects, critical situations, etc..). My lack of time is the main reason the bot is still up :).
In closing - Holiday time are among us - for the USA, Thanksgiving is past - and for many of us the year-end holidays come quickly. I wish you all the very best holidays and other year-ends activities.