AFS operation meeting 2015-11-16
present: Dan, Massimo, Kuba, Jan
Issues:
- afs286 got stuck, needed hard restart (plus long salvage time). BOINC'ed? (work.boinc lives there, nowadays; might have been inaccessible already a few days before)
- better / faster ways to detect this? various symptoms (user tickets, swap full, high-load on afscron)
- 1.6.15 crash - no progress, looks like bad merge 1.6.7->1.6.9
- observed "idle threads = 1" on a fileservers == stuck?
- ABS logrotate: added, but should be more than 7 days -> 28 days?
There are minutes attached to this event.
Show them.