Project

General

Profile

Actions
Copy link

Bug #2536

closed

librados crashed while getting stat of an object

Added by Xiaopong Tran almost 12 years ago. Updated over 11 years ago.

Status:
Can't reproduce
Priority:
Normal
Assignee:
-
Category:
librados
Target version:
-
% Done:

0%

Source:
Community (user)
Tags:
Backport:
Regression:
Severity:
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

librados crashed while getting stat of an object:

./log/SubsystemMap.h: In function 'bool ceph::log::SubsystemMap::should_gather(unsigned int, int)' thread 7f52adaf3700 time 2012-06-11 18:33:47.455897
./log/SubsystemMap.h: 74: FAILED assert(sub < m_subsys.size())
ceph version 0.47.2 (commit:8bf9fde89bd6ebc4b0645b2fe02dadb1c17ad372)
1: (()+0xebbc0) [0x7f5298a92bc0]
2: (librados::IoCtxImpl::stat(object_t const&, unsigned long*, long*)+0x597) [0x7f5298ab2a87]
3: (rados_stat()+0x3a) [0x7f5298a975fa]
4: (x_stat(enif_environment_t*, int, unsigned long const*)+0x1cd) [0x7f5298f11f65]
5: (process_main()+0x4f32) [0x544122]
6: /usr/local/erlang/lib/erlang/erts-5.9.1/bin/beam.smp() [0x4a7d08]
7: /usr/local/erlang/lib/erlang/erts-5.9.1/bin/beam.smp() [0x5bcb20]
8: (()+0x7e9a) [0x7f52b378be9a]
9: (clone()+0x6d) [0x7f52b32b14bd]

NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.
terminate called after throwing an instance of 'ceph::FailedAssertion'
Aborted (core dumped)

Attached is the objdump file.

Files

Download all files
objdump.txt (1.24 MB) objdump.txt Xiaopong Tran, 06/11/2012 04:01 AM
objdump.txt (8.24 MB) objdump.txt Benjamin Schulz, 11/25/2012 06:36 PM
Actions
Copy link
 #1

Updated by Sage Weil almost 12 years ago

Have you seen this problem since then? It looks like it could be due to racing with rados startup or shutdown...

Actions
Copy link
 #2

Updated by Sage Weil almost 12 years ago

  • Status changed from New to Need More Info
Actions
Copy link
 #3

Updated by Sage Weil over 11 years ago

  • Priority changed from High to Normal
Actions
Copy link
 #4

Updated by Sage Weil over 11 years ago

  • Status changed from Need More Info to Can't reproduce
Actions
Copy link
 #5

Updated by Benjamin Schulz over 11 years ago

Hi,

I got the same assertion:

radosgw-admin user create
./log/SubsystemMap.h: In function 'bool ceph::log::SubsystemMap::should_gather(unsigned int, int)' thread 7f8fa8546760 time 2012-11-26 02:41:55.267779
./log/SubsystemMap.h: 62: FAILED assert(sub < m_subsys.size())
ceph version 0.48.1argonaut (a7ad701b9bd479f20429f19e6fea7373ca6bba7c)
1: (()+0x17c3a7) [0x7f8fa770f3a7]
2: (MonClient::build_initial_monmap()+0x1e9) [0x7f8fa77da249]
3: (librados::RadosClient::connect()+0x48) [0x7f8fa77231b8]
4: (RGWRados::initialize()+0x49) [0x48ece9]
5: (RGWCache<RGWRados>::initialize()+0x17) [0x4a31e7]
6: (RGWRados::init_storage_provider(CephContext*)+0x30) [0x48ebe0]
7: (main()+0xfd0) [0x42c8c0]
8: (__libc_start_main()+0xfd) [0x7f8fa5cbcead]
9: radosgw-admin() [0x432761]
NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this.
  • Caught signal (Segmentation fault) *
    in thread 7f8fa8546760
    ceph version 0.48.1argonaut (a7ad701b9bd479f20429f19e6fea7373ca6bba7c)
    1: radosgw-admin() [0x43d8b2]
    2: (()+0xf030) [0x7f8fa715c030]
    3: (ceph::__ceph_assert_fail(char const    , char const*, int, char const*)+0x38f) [0x445edf]
    4: (()+0x17c3a7) [0x7f8fa770f3a7]
    5: (MonClient::build_initial_monmap()+0x1e9) [0x7f8fa77da249]
    6: (librados::RadosClient::connect()+0x48) [0x7f8fa77231b8]
    7: (RGWRados::initialize()+0x49) [0x48ece9]
    8: (RGWCache<RGWRados>::initialize()+0x17) [0x4a31e7]
    9: (RGWRados::init_storage_provider(CephContext*)+0x30) [0x48ebe0]
    10: (main()+0xfd0) [0x42c8c0]
    11: (__libc_start_main()+0xfd) [0x7f8fa5cbcead]
    12: radosgw-admin() [0x432761]
    Segmentation fault

I'm running v0.48.1 on debian wheezy. The system is setup in two VMs, I can reproduce it every time. Contact me, if you're interested in the VM-Images.

best Regards
-- Benjamin

Actions
Copy link
 #6

Updated by Greg Farnum over 11 years ago

Hey Benjamin, this is the same assert but quite a different call chain — can you create a new bug? Preferably with some logs of the crash?

Actions
Copy link

Also available in: Atom PDF