Hello folks, Ran into this one today on a machine using kRBD. Completely locked the machine: Apr 1 17:35:10 nfs1 kernel: [492418.251665] Call Trace: Apr 1 17:35:10 nfs1 kernel: [492418.251675] [<ffffffffa02086b4>] reset_changed_osds+0x74/0xa0 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251682] [<ffffffffa020b0c2>] ceph_osdc_handle_map+0x212/0x3e0 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251689] [<ffffffffa0207547>] dispatch+0xa7/0x120 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251694] [<ffffffffa0201045>] process_message+0xa5/0xc0 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251700] [<ffffffffa0204fa1>] try_read+0x2e1/0x440 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251705] [<ffffffffa0205100>] ? try_read+0x440/0x440 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251711] [<ffffffffa0205192>] con_work+0x92/0x1c0 [libceph] Apr 1 17:35:10 nfs1 kernel: [492418.251716] [<ffffffff81071baa>] process_one_work+0x11a/0x480 Apr 1 17:35:10 nfs1 kernel: [492418.251720] [<ffffffff81072bc5>] worker_thread+0x165/0x370 Apr 1 17:35:10 nfs1 kernel: [492418.251723] [<ffffffff81072a60>] ? manage_workers.isra.29+0x130/0x130 Apr 1 17:35:10 nfs1 kernel: [492418.251727] [<ffffffff81077b63>] kthread+0x93/0xa0 Apr 1 17:35:10 nfs1 kernel: [492418.251733] [<ffffffff816a3ee4>] kernel_thread_helper+0x4/0x10 Apr 1 17:35:10 nfs1 kernel: [492418.251736] [<ffffffff81077ad0>] ? flush_kthread_worker+0xb0/0xb0 Apr 1 17:35:10 nfs1 kernel: [492418.251739] [<ffffffff816a3ee0>] ? gs_change+0x13/0x13 Anybody run into this before? I'm running kernel 3.5.7.2 I came across a few patches that touch relevant functions: https://patchwork.kernel.org/patch/1913871/ >From what i can tell, this can pulled into 3.4.26 and 3.7.3. I can't tell if it went into 3.5.x at all. Is this a known bug that's already been fixed? This happened while I was doing a rolling upgrade of the Ceph cluster -- doing a "service ceph restart osd.$x" for 60 OSDs. 30 second sleep in between each to allow things to settle. - Travis -- 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
