Looks like these are all sockets that don't get closed because we send the associated Pipes into reconnect, but they never actually do so (and never move on out of it into shutting down). We made some changes to how reconnect and cleanup are done that probably weren't complete/missed a case. (Maybe only clients actually try to reconnect, but everybody goes into that state?) For tomorrow! On Mon, Dec 5, 2011 at 11:36 AM, Tommi Virtanen <tommi.virtanen@xxxxxxxxxxxxx> wrote: > On Mon, Dec 5, 2011 at 11:05, Noah Watkins <jayhawk@xxxxxxxxxxx> wrote: >> root@issdm-23:~# lsof -n -p 1394 >> COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME >> ceph-mds 1394 root 12u sock 0,6 0t0 11560 can't >> identify protocol > > There's 506 of these guys in the listing. I would say there's a leak > of sockets that get created with socket(2), but don't get bind(2) or > connect(2) called on them. > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
