On Thu, May 07, 2026 at 05:18:14PM +0200, Mark Wielaard wrote: > Lets discuss the previous year and the goals and budget for the > upcoming year. The Sourceware (financial) year runs from/to May 15 > which is when we joined the Software Freedom Conservancy as a member > project. https://sfconservancy.org/news/2023/may/15/sourceware-joins-sfc/
For those that couldn't attend. A summary of what was discussed. Thanks to SFC staff for attending and guidance on these budget talks and for providing a stable financial administrative base. Discussion has two goals. - Prepare for yearly report, what did we predict we would spend/raise and what did we actually spend/raise? - And what do we believe the community wants the Sourceware PLC to focus on next year and does that involve spending and raising more money? First is kind of easy, we look at the yearly report from last year. https://inbox.sourceware.org/[email protected] "Leaving us with $10,095.15 at the end of the year" "The PLC wants to use some of our current budget to add a bigger machine... and help with our goal to isolate more services on separate machines or VMs." Doesn't say how much, but we hoped to not spend all of it, keep ~half for emergencies and try to raise enough money in ~3 years to do another hardware refresh. How did we do? Not badly! We finished the year with $10,017.59. Spending $6,358.98 and raising $6,332.12. server1 cost us $6,195.31 (including tax and shipping). That purchase was really excellently timed. The 1.5TB ram would now cost as much as the whole server. Other expenses were paypal/banking fees and domain fees. And with only individual donations. Last year the donations were ~$250 a month, this year they doubled to ~$500 a month. We currently don't attach a dollar amount to the "in kind" donations of hosting and hands on admin help from Red Hat and OSUOSL. Red Hat used to host 2, now 3 machines (and for small hardware fixes they sometimes just provide the parts). And OSUOSL used to host two small 1u servers for us, now a much bigger 2u machine. We are also using some "cloud" servers from both of them. OSUOSL said they like us to pay something for the colocation plus bandwith, but aren't insisting. Given the current improved financial situation, the recommendation to the PLC is to talk to the OSL to see what is fair. With their help our services all moved into VMs. Our hardware, servers, VM setup looks pretty nicely now: https://sourceware.org/sourceware-wiki/servers-and-services-2026 This should definitely get us going for the next 2 to 3 years. So for next year we can focus more on services and helping the admins out instead of spending on actual infrastructure/hosting/hardware. We could try to be a little more ambitious and diversify our corporate funding and approach some grant makers, so we rely less heavily on individual donations. Which brings us to the tricky question, what does the community want us to focus on and how much money can we dedicate to that (if we don't just want to depend on volunteer admins)? See the survey and our previous discussions: https://inbox.sourceware.org/[email protected] The majority of people didn't vote or said everything was fine. But there is a clear "winner". Forgejo process improvements. We should make a list of upstream bug/issues that are blocking us. For example "restricted accounts" aren't really working out for us. And user/account administration/rights should be more flexible so it doesn't require full admin rights. It also would be nice to have very light weight accounts that can just do agit. Recommend the PLC set aside some money to get such upstream issues fixed to help forge admins. And ask Claudio and Richard for guidance. There are several plans that would go faster if we could hire some consultant or a sysadmin that could execute them. https://sourceware.org/sourceware-security-vision.html#plans Specifically moving specific services into VMs, upgrading their versions, and automating their administration/updates using (e.g.) Ansible would be useful. This could be split up for specific services. Specifically for bugzilla, buildbot, patchwork and public-inbox. Recommend the PLC look into hiring consultants to accellerate this work, if enough additional funding can be found. Next up, finalizing the Sourceware year report and the budget writeup for the funding drive.
