Actions
Bug #40698
closedmgr: openSSL SSL_library_init() failed to initialize
Status:
Resolved
Priority:
Urgent
Assignee:
Category:
ceph-mgr
Target version:
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Description
Recently, when trying to do some test, the ceph-mgr failed to boot:
2019-07-09T13:29:45.164+0800 7ff56b7e3b80 -1 /home/xuxuehan/mnt/github-ceph/src/common/lockdep.cc: In function 'int lockdep_will_lock(const char*, int, bool, bool)' thread 7ff56b7e3b80 time 2019-07-09T13:29:45.155129+0800 /home/xuxuehan/mnt/github-ceph/src/common/lockdep.cc: 305: ceph_abort_msg("abort() called") ceph version v15.0.0-2554-g221d976e3c (221d976e3c21e0d99c82f3d808b7bd1f7e5a406d) octopus (dev) 1: (ceph::__ceph_abort(char const*, int, char const*, std::string const&)+0xfe) [0x7ff5623ea874] 2: (lockdep_will_lock(char const*, int, bool, bool)+0x44b) [0x7ff5624d1b6f] 3: (ceph::mutex_debug_detail::mutex_debugging_base::_will_lock(bool)+0x4a) [0x7ff5624dd89c] 4: (ceph::shared_mutex_debug::lock()+0x3a) [0x7ff56257bbee] 5: (()+0x143a532) [0x7ff56241b532] 6: (CRYPTO_mem_ctrl()+0x39) [0x7ff55fdd1be9] 7: (()+0x4a60e) [0x7ff54a51b60e] 8: (SSL_COMP_get_compression_methods()+0x9) [0x7ff54a51d899] 9: (SSL_library_init()+0x179) [0x7ff54a5237b9] 10: (init_hashlib()+0x1f) [0x7ff54a74614f] 11: (_PyImport_LoadDynamicModule()+0x99) [0x7ff560d0ed69] 12: (()+0xf7e41) [0x7ff560d0ce41] 13: (()+0xf808d) [0x7ff560d0d08d] 14: (PyImport_ImportModuleLevel()+0x15e) [0x7ff560d0da6e] 15: (()+0xdbd1f) [0x7ff560cf0d1f] 16: (PyObject_Call()+0x43) [0x7ff560c60a63] 17: (PyEval_CallObjectWithKeywords()+0x47) [0x7ff560cf28f7] 18: (PyEval_EvalFrameEx()+0x46b5) [0x7ff560cf75b5] 19: (PyEval_EvalCodeEx()+0x7ed) [0x7ff560cfc03d]
Files
Updated by Sebastian Wagner almost 5 years ago
- Subject changed from mgr: failed to boot to mgr: openSSL SSL_library_init() failed to initialize
Could you please attach the full mgr logs?
Updated by Radoslaw Zarzynski almost 5 years ago
Could we have a log with lockdep_force_backtrace set to true?
The idea is to figure out the place the where the lock has been taken initially.
Updated by Kefu Chai almost 5 years ago
- Status changed from New to Need More Info
- Assignee set to Xuehan Xu
Updated by Xuehan Xu almost 5 years ago
Radoslaw Zarzynski wrote:
Could we have a log with lockdep_force_backtrace set to true?
The idea is to figure out the place the where the lock has been taken initially.
Here is the log with "lockdep_force_backtrace = true"
Updated by Sebastian Wagner almost 5 years ago
- Has duplicate Bug #40724: mgr: recursive lock in vstart added
Updated by Sebastian Wagner almost 5 years ago
- Priority changed from Normal to Urgent
Updated by Sebastian Wagner almost 5 years ago
- Related to Bug #40688: mgr/dashboard: ceph dashboard tests Jenkins job fails with "*** Caught signal (Aborted) ** in thread 7f9d14518b80 thread_name:ceph-mgr" added
Updated by Radoslaw Zarzynski almost 5 years ago
- Status changed from Need More Info to In Progress
- Assignee changed from Xuehan Xu to Radoslaw Zarzynski
Updated by Radoslaw Zarzynski almost 5 years ago
- Status changed from In Progress to Fix Under Review
Updated by Lenz Grimmer almost 5 years ago
Radoslaw Zarzynski wrote:
https://gist.github.com/rzarzynski/b2e73a4fefd86c1cbbafc76f8e3ee55f
Thanks for the detailed analysis! Impressive :)
Updated by Sage Weil almost 5 years ago
- Status changed from Fix Under Review to Resolved
Updated by Laura Paduano almost 5 years ago
- Related to Bug #40813: mgr/dashboard: Remove --nolockdep option from e2e script and vstart_runner.py added
Actions