[MXS-1912] Maxscale takes signal 11 w/backtrace and ends up hung Created: 2018-06-08  Updated: 2018-06-11  Resolved: 2018-06-11

Status: Closed
Project: MariaDB MaxScale
Component/s: maxadmin
Affects Version/s: 2.2.5
Fix Version/s: 2.2.7

Type: Bug Priority: Major
Reporter: Richard Lane Assignee: Unassigned
Resolution: Duplicate Votes: 0
Labels: None
Environment:

RHEL7.4 container

Attachments: Text File maxscale-hang.log    

 Description   

In the attached log there are two problems found:
1. Two cases where maxscale received a signal 11 and generated a stack trace and was restarted
2. This log was saved when maxscale was apparently hung. There were a number of maxadmin list server commands hung (which we do to verify sanity of maxscale). Running a maxadmin list server just hangs.

Once we killed all the hung maxadmin list server processes and restarted maxscale, everything was back to normal.

So two questions:
(a) why is maxscale taking a signal 11 and restarting
(b) why is maxscale hanging (unfortunately this is container environment so there is no backtrace capability).

 Comments   
Comment by Johan Wikman [ 2018-06-11 ]

This is the same bug as MXS-1875 that has been fixed in MaxScale 2.2.7.

The errors

2018-06-05 14:46:49   error  : Failed to write message: 11, Resource temporarily unavailable
 
2018-06-05 14:46:49   error  : Failed to write message: 11, Resource temporarily unavailable

that are present in the log are related to MXS-1878 , which is also fixed by MaxScale 2.2.7.

Comment by Johan Wikman [ 2018-06-11 ]

Duplicate of MXS-1875 (and MXS-1878)

Generated at Thu Feb 08 04:10:19 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.