Description of problem: nfs process called due to SIGSEGV. the glusterfs source was built with lefence. Program received signal SIGSEGV, Segmentation fault. 0x00007f39d6372c9b in afr_frame_return (frame=0x7f39d8055b08) at afr-common.c:968 968 call_count = --local->call_count; (gdb) p local $1 = (afr_local_t *) 0x0 Version-Release number of selected component (if applicable): glusterfs-3.3.30qa19 How reproducible: always Steps to Reproduce: 1. Create a strip-rep or dist-stripe-rep volume. 2. mount the volume from one machine via fuse and from another machine via nfs. 3. Untar the linux kernel from both the clients in different locations. 4. start dbench from both the clients from different directories. Actual results: nfs process carshed due to SIGSEGV. (gdb) bt #0 0x00007f39d6372c9b in afr_frame_return (frame=0x7f39d8055b08) at afr-common.c:968 #1 0x00007f39d63430f7 in afr_openfd_fix_open_cbk (frame=0x7f39d8055b08, cookie=0x0, this=0x7f39d66716b8, op_ret=-1, op_errno=2, fd=0x7f39d388a038) at afr-open.c:324 #2 0x00007f39d66b30c4 in client3_1_opendir_cbk (req=0x7f39d1aaa12c, iov=0x7f39d1aaa16c, count=1, myframe=0x7f39d8050304) at client3_1-fops.c:2154 #3 0x00007f39d94997b0 in rpc_clnt_handle_reply (clnt=0x7f39d2b5dee8, pollin=0x7f39ca9bafb0) at rpc-clnt.c:789 #4 0x00007f39d9499b37 in rpc_clnt_notify (trans=0x7f39d2b67d40, mydata=0x7f39d2b5df18, event=RPC_TRANSPORT_MSG_RECEIVED, data=0x7f39ca9bafb0) at rpc-clnt.c:908 #5 0x00007f39d9495d14 in rpc_transport_notify (this=0x7f39d2b67d40, event=RPC_TRANSPORT_MSG_RECEIVED, data=0x7f39ca9bafb0) at rpc-transport.c:498 #6 0x00007f39d367924d in socket_event_poll_in (this=0x7f39d2b67d40) at socket.c:1675 #7 0x00007f39d36797d1 in socket_event_handler (fd=20, idx=11, data=0x7f39d2b67d40, poll_in=1, poll_out=0, poll_err=0) at socket.c:1790 #8 0x00007f39d96ea808 in event_dispatch_epoll_handler (event_pool=0x7f39d8fadf60, events=0x7f39d9031374, i=3) at event.c:794 #9 0x00007f39d96eaa2b in event_dispatch_epoll (event_pool=0x7f39d8fadf60) at event.c:856 #10 0x00007f39d96eadb6 in event_dispatch (event_pool=0x7f39d8fadf60) at event.c:956 #11 0x0000000000407ace in main (argc=7, argv=0x7fff6e596228) at glusterfsd.c:1601 Expected results: process should not crash. Additional info: I have archived the all log files.
CHANGE: http://review.gluster.com/2687 (cluster/afr: Stack wind with correct frame) merged in master by Vijay Bellur (vijay)
Verified with stripe-rep volume and dbench completed successfully. No crashes found.