VDI check crash with erasure coding and not enouch nodes

Bug #1335865 reported by sirio81
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
sheepdog
Undecided
Unassigned

Bug Description

This is a cluster with 3 nodes and formatted with -c 2:2.
Importing a qcow2 works fine but vdi check fails:

time qemu-img convert -f qcow2 test.qcow2 sheepdog:test; time dog vdi check test

  0.0 % [> ] 0.0 MB / 10 GB fixed missing 7c2b2500000000, copy index 2
dog exits unexpectedly (Segmentation fault).
dog.c:373: crash_handler
/lib/x86_64-linux-gnu/libpthread.so.0(+0xf02f) [0x7f9d0d2ab02f]
/lib/x86_64-linux-gnu/libc.so.6(cfree+0x18) [0x7f9d0cd73ac8]
vdi.c:1665: check_erasure_object
work.c:303: worker_thread_request_done
event.c:210: do_event_loop
common.c:288: work_queue_wait
vdi.c:1836: do_vdi_check
vdi.c:1857: vdi_check
dog.c:565: main
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xfc) [0x7f9d0cd16eac]
dog() [0x404068]
Segmentation fault

dog cluster info
Cluster status: running, auto-recovery enabled
Cluster created at Mon Jun 30 15:51:26 2014
Epoch Time Version
2014-06-30 15:51:26 1 [192.168.10.4:7000, 192.168.10.5:7000, 192.168.10.6:7000]

dog vdi list
  Name Id Size Used Shared Creation time VDI id Copies Tag
  test 0 10 GB 10 GB 0.0 MB 2014-06-30 15:51 7c2b25 2:2

No errors are reported in sheep.log

Jun 30 15:29:45 INFO [main] md_add_disk(338) /mnt/sheep/0, vdisk nr 217, total disk 1
Jun 30 15:29:45 NOTICE [main] get_local_addr(519) found IPv4 address
Jun 30 15:29:45 INFO [main] send_join_request(828) IPv4 ip:192.168.10.4 port:7000
Jun 30 15:29:45 NOTICE [main] nfs_init(600) nfs server service is not compiled
Jun 30 15:29:45 INFO [main] check_host_env(493) Allowed open files 1024000, suggested 6144000
Jun 30 15:29:45 INFO [main] main(942) sheepdog daemon (version 0.8.0_219_g49857cb) started
Jun 30 15:29:48 INFO [main] zk_find_master(836) detect master leave, start to compete master
Jun 30 15:29:48 INFO [main] zk_find_master(836) detect master leave, start to compete master
Jun 30 15:29:48 INFO [main] zk_find_master(836) detect master leave, start to compete master
Jun 30 15:32:06 INFO [main] rx_main(811) req=0x349d2a0, fd=23, client=::1:35354, op=NEW_VDI, data=(not string)
Jun 30 15:32:06 INFO [main] tx_main(863) req=0x349d2a0, fd=23, client=::1:35354, op=NEW_VDI, result=00
Jun 30 15:40:26 INFO [main] rx_main(811) req=0x7f56d4000960, fd=23, client=127.0.0.1:39736, op=MAKE_FS, data=(not string)
Jun 30 15:40:26 INFO [main] tx_main(863) req=0x7f56d4000960, fd=23, client=127.0.0.1:39736, op=MAKE_FS, result=00
Jun 30 15:40:38 INFO [main] rx_main(811) req=0x7f56d4000d10, fd=23, client=::1:35379, op=NEW_VDI, data=(not string)
Jun 30 15:40:38 INFO [main] tx_main(863) req=0x7f56d4000d10, fd=23, client=::1:35379, op=NEW_VDI, result=00
Jun 30 15:51:26 INFO [main] rx_main(811) req=0x34b6c70, fd=23, client=127.0.0.1:39759, op=MAKE_FS, data=(not string)
Jun 30 15:51:26 INFO [main] tx_main(863) req=0x34b6c70, fd=23, client=127.0.0.1:39759, op=MAKE_FS, result=00
Jun 30 15:51:41 INFO [main] rx_main(811) req=0x34bca80, fd=23, client=::1:35402, op=NEW_VDI, data=(not string)
Jun 30 15:51:41 INFO [main] tx_main(863) req=0x34bca80, fd=23, client=::1:35402, op=NEW_VDI, result=00

Revision history for this message
sirio81 (sirio81) wrote :

It affects stable 0.82 and master 0.8.0_219_g49857cb.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers