Actions
Bug #59562
closed[pacific] segfault on FIPS enabled server at "radosgw-admin period update" command
% Done:
0%
Source:
Community (dev)
Tags:
Backport:
Regression:
No
Severity:
2 - major
Reviewed:
Affected Versions:
ceph-qa-suite:
rgw
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
After https://tracker.ceph.com/issues/52900 PR merged we tried to enable RGW on FIPS enabled host and got the following error:
-25> 2023-04-26T13:44:06.080+0000 7fe277b18600 5 monclient: authenticate success, global_id 10611936 -24> 2023-04-26T13:44:06.080+0000 7fe277b18600 10 monclient: _renew_subs -23> 2023-04-26T13:44:06.080+0000 7fe277b18600 10 monclient: _send_mon_message to mon.a at v2:<IP>:3300/0 -22> 2023-04-26T13:44:06.080+0000 7fe24afe5640 4 set_mon_vals no callback set -21> 2023-04-26T13:44:06.080+0000 7fe277b18600 10 monclient: _renew_subs -20> 2023-04-26T13:44:06.080+0000 7fe277b18600 10 monclient: _send_mon_message to mon.a at v2:<IP>:3300/0 -19> 2023-04-26T13:44:06.080+0000 7fe277b18600 1 librados: init done -18> 2023-04-26T13:44:06.080+0000 7fe277b18600 5 asok(0x5610d50b4440) register_command cr dump hook 0x5610d5212f18 -17> 2023-04-26T13:44:06.080+0000 7fe27570e640 10 monclient: get_auth_request con 0x5610d521d6a0 auth_method 0 -16> 2023-04-26T13:44:06.084+0000 7fe2497e2640 4 mgrc handle_mgr_map Got map version 190 -15> 2023-04-26T13:44:06.084+0000 7fe2497e2640 4 mgrc handle_mgr_map Active mgr is now [v2:<IP>:6800/44795856,v1:<IP>:6801/44795856] -14> 2023-04-26T13:44:06.084+0000 7fe2497e2640 4 mgrc reconnect Starting new session with [v2:<IP>:6800/44795856,v1:<IP>:6801/44795856] -13> 2023-04-26T13:44:06.084+0000 7fe275f0f640 10 monclient: get_auth_request con 0x5610d5220f70 auth_method 0 -12> 2023-04-26T13:44:06.084+0000 7fe27570e640 10 monclient: get_auth_request con 0x7fe1c80528d0 auth_method 0 -11> 2023-04-26T13:44:06.088+0000 7fe274f0d640 10 monclient: get_auth_request con 0x5610d52246e0 auth_method 0 -10> 2023-04-26T13:44:06.092+0000 7fe275f0f640 10 monclient: get_auth_request con 0x5610d5227f70 auth_method 0 -9> 2023-04-26T13:44:06.100+0000 7fe27570e640 10 monclient: get_auth_request con 0x7fe26802a070 auth_method 0 -8> 2023-04-26T13:44:06.100+0000 7fe274f0d640 10 monclient: get_auth_request con 0x7fe2680389a0 auth_method 0 -7> 2023-04-26T13:44:06.104+0000 7fe275f0f640 10 monclient: get_auth_request con 0x7fe26c076c10 auth_method 0 -6> 2023-04-26T13:44:06.108+0000 7fe27570e640 10 monclient: get_auth_request con 0x7fe26801b3f0 auth_method 0 -5> 2023-04-26T13:44:06.112+0000 7fe274f0d640 10 monclient: get_auth_request con 0x7fe26804b3f0 auth_method 0 -4> 2023-04-26T13:44:06.116+0000 7fe275f0f640 10 monclient: get_auth_request con 0x7fe26c0aa040 auth_method 0 -3> 2023-04-26T13:44:06.120+0000 7fe27570e640 10 monclient: get_auth_request con 0x7fe2600082f0 auth_method 0 -2> 2023-04-26T13:44:06.124+0000 7fe274f0d640 10 monclient: get_auth_request con 0x7fe26c0ad0b0 auth_method 0 -1> 2023-04-26T13:44:06.128+0000 7fe275f0f640 10 monclient: get_auth_request con 0x7fe27003e040 auth_method 0 0> 2023-04-26T13:44:06.140+0000 7fe277b18600 -1 *** Caught signal (Segmentation fault) ** in thread 7fe277b18600 thread_name:radosgw-admin ceph version 16.2.11 (3cf40e2dca667f68c6ce3ff5cd94f01e711af894) pacific (stable) 1: /lib64/libc.so.6(+0x54df0) [0x7fe279cb4df0] --- logging levels --- ... -2/-2 (syslog threshold) 99/99 (stderr threshold) --- pthread ID / name mapping for recent threads --- 140609872340544 / ms_dispatch 140609897518656 / io_context_pool 140610350593600 / ms_dispatch 140610375771712 / io_context_pool 140610384164416 / fn_anonymous 140610601276992 / msgr-worker-2 140610609669696 / msgr-worker-1 140610618062400 / msgr-worker-0 140610647459328 / radosgw-admin max_recent 500 max_new 500
We made gdb run, threads info and backtraces could be found in attachments. This issue blocks using Ceph RGW on FIPS enabled cluster.
Files
Updated by Casey Bodley 12 months ago
- Related to Bug #52900: segfault on FIPS enabled server as result of EVP_md5 disabled in openssl added
Updated by Mark Kogan 12 months ago
@Peter Razumovsky
May ask which OS please for repro? (there are differences between openssl v1 and v3)
Updated by Mark Kogan 12 months ago
reproduced, working on backport of https://tracker.ceph.com/issues/58332 to pacific at https://tracker.ceph.com/issues/59579
Updated by Mark Kogan 12 months ago
- Is duplicate of Bug #58332: On version 17.2.5-8.el9cp, "Segmentation fault" while uploading object(regular/multipart) on FIPS enabled cluster. added
Updated by Mark Kogan 12 months ago
backport PR to pacific:
https://github.com/ceph/ceph/pull/51266
Updated by Casey Bodley 12 months ago
- Status changed from In Progress to Fix Under Review
Updated by Peter Razumovsky 12 months ago
Sure, it's ubuntu 20.04 with kernel 5.4.0
Updated by Casey Bodley 11 months ago
- Status changed from Fix Under Review to Resolved
Actions