Good news everyone.
After a month, the old server managed to boot (in fact, the RAID controller was not dead, but just too stupid to kick out the broken disk). The colo facility pulled the broken disk, so it should work again for some time if nothing else breaks.
Next nightly build is scheduled at 17:41 CET, as usual, so starts in about 20 minutes. Progress can be seen on the other Jenkins.
New box will still be rolled out at some point, but is delayed because of a public holiday and shipping delays (enterprise class SSDs are hard to get), which should not be that much of a problem now though.