I've been upgrading the machines to 3.2.0-0.bpo.2 (3.2.17), and rescheduled the checkarray to occur every weekend. So far, after two weekends, no more freeze occurrences. Previously, every 1-2 checkarray runs the machine used to stall. Seems the bug is fixed now.
-- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org