Am 06.12.2015 um 10:29 schrieb Hermès BÉLUSCA - MAÏTO: > Anybody really knows what happened there ? (since it’s not the first > time this part of the infra fails like that).
This time, we really had a major power failure at Jan's site that lasted for two days. The UPS systems were able to keep up most of the systems for 10 more minutes till these were gracefully shut down. Actually, this has been the only time we had such a major power outage. Nevertheless, there is a single server "Fezile" from 2007 (responsible for iso.reactos.org, Doxygen, VMware Testbot) that has been troubling over the year: * In December 2014, its motherboard failed. As a replacement one was cheaply available, we didn't need more performance for its tasks and the system was fully set up, it was decided to just order that one and continue using the system. Unfortunately, replacing it took longer than expected. It's now fully up and running again, but next time we will definitely order a new server. * Recently, we also had Fezile crashing several times due to a faulty UPS it was connected to. This has been fixed in the meantime. * The latest total power outage is of course not Fezile's fault, but affected each and every server there. Nevertheless, those other servers (e.g. additional Testbots) have basically been up every day of the year, which I find quite reliable. I'm open for every affordable idea to make this infrastructure more reliable. I just personally don't think we can make better than Jan's site, since it offers us total flexibility on what servers to install, how to interconnect them, and the help of a definite ReactOS supporter in case something fails. Cheers, Colin
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ Ros-dev mailing list Ros-dev@reactos.org http://www.reactos.org/mailman/listinfo/ros-dev
