Support #38374
Updated by Patrick Donnelly about 5 years ago
It seems when Docker creates the first device with dmsetup, the MDS crashes and all further MDS replaying the log crash too. Mounted with: <pre> 10.10.0.16,10.10.0.17,10.10.0.18:/gitlab/runner/docker on /var/lib/docker type ceph (rw,noatime,name=gitlabrunner,secret=<hidden>,acl,wsize=16777216,write_congestion_kb=127616) </pre> The mount point has xattr set to use a different size=1 pool and an alternate namespace if it matters (same as Kubernetes does in another pool). <pre> Thread 25 "fn_anonymous" received signal SIGABRT, Aborted. [Switching to Thread 0x7fffe07fc700 (LWP 32514)] __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51 51 ../sysdeps/unix/sysv/linux/raise.c: No such file or directory. (gdb) bt #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51 #1 0x00007fffede11801 in __GI_abort () at abort.c:79 #2 0x00007fffef401790 in ceph::__ceph_assert_fail (assertion=<optimized out>, file=<optimized out>, line=<optimized out>, func=<optimized out>) at /home/jerome.poulin/GIT/ceph/src/common/assert.cc:66 #3 0x00007fffef401807 in ceph::__ceph_assert_fail (ctx=...) at /home/jerome.poulin/GIT/ceph/src/common/assert.cc:71 #4 0x00005555557ba700 in virtual thunk to C_IO_MDC_TruncateFinish::finish(int) () #5 0x00005555558da4e9 in Context::complete (r=-1, this=0x55555689a820) at /home/jerome.poulin/GIT/ceph/src/include/Context.h:77 #6 MDSIOContextBase::complete (this=0x55555689a820, r=-1) at /home/jerome.poulin/GIT/ceph/src/mds/MDSContext.cc:116 #7 0x00007fffef3ffbde in Finisher::finisher_thread_entry (this=0x5555564c3980) at /home/jerome.poulin/GIT/ceph/src/common/Finisher.cc:68 #8 0x00007fffeed0c6db in start_thread (arg=0x7fffe07fc700) at pthread_create.c:463 #9 0x00007fffedef288f in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95 </pre>