Custom Query (1145 matches)
Results (313 - 315 of 1145)
Ticket | Resolution | Summary | Owner | Reporter |
---|---|---|---|---|
#978 | fixed | Possible installation hang with lvremove | geofft | |
Description |
I ran into the lvremove -f /dev/athena/keep_2 from install-debathena.sh hanging on lola-granola. I was attempting to reinstall it after an abortive previous install. We should see if this is reproducible when installing on a machine that has already been partially (or wholly) installed. Here's the debugging info I got: debathena / personal / geofft 17:23 (Geoffrey Thomas) .... ... .. why is `lvremove -f /dev/athena/keep_2` hanging in sys_semtimedop? debathena / personal / geofft 17:24 (Geoffrey Thomas) with no other threads? debathena / personal / geofft 17:57 (Geoffrey Thomas) According to gdb, we're in semop in dm_udev_wait in ?? in lv_deactivate in ?? in ?? in lock_vol in lv_remove_single debathena / personal / geofft 18:12 (Geoffrey Thomas) Also at about the same time I got udevd-work[18528]: inotify_add_watch(6, /dev/dm-2, 10) failed: No such file or directory debathena / personal / geofft 18:12 (Geoffrey Thomas) And strace says semop(0, {{0, 0, 0}}, 1 If we can't reproduce this and don't see this again, this ticket should be closed. |
|||
#847 | fixed | Port config-package-dev to Debhelper 7/8 | geofft | geofft |
Description |
Debhelper 7 gained a lot of the cool stuff that people used CDBS for. If it's possible for something config-package-dev like to exist in Debhelper, then there's a strong popularity argument to moving our code to Debhelper + dh_configpackage. |
|||
#1478 | fixed | policy-rc.d in reactivate gets scribbled over by d-i, and is also obsolete | jdreed | |
Description |
There are a couple of problems here: First, /usr/sbin/policy-rc.d is supposed to be managed by update-alternatives, per the documentation. That having been said, it doesn't matter, because d-i will happily scribble over it anyway. It is unclear to me that we still need a policy-rc.d outside the chroot. I believe the primary problem here was CUPS, which we had to ensure was running inside the chroot, and never outside it. I think that's no longer true, because we've given up on BrowsePolling?. Adding a custom printer to CUPS, for the duration of your chroot login, should not require restarting the daemon. So we should rip out the CUPS restart code, and the chroot itself should be given a simple policy-rc.d that exits 101, but there's no need for it to exist outside the chroot. Thoughts? Barring objections, I will submit a PR for this tomorrow. |