Issues with lupus (file system for irma) closed
We were noticed of lupus (the file system for irma) being slow on some nodes and traced it to an issue with a specific disk target. While resolving that we rebooted one of the storage servers and it took somewhat longer than expected for HA services to act to provide the targets affected.
As far as we can tell, this is resolved now. In theory, it should not have affected ongoing I/O (except for the delay), but there can always be bugs and such, so it may even have lead to failed jobs.