[MCOL-261] If DBRMControllerNode crashes, system remains in not ready state Created: 2016-08-04  Updated: 2016-09-16  Resolved: 2016-09-16

Status: Closed
Project: MariaDB ColumnStore
Component/s: MariaDB Server
Affects Version/s: 1.0.1
Fix Version/s: 1.0.3

Type: Bug Priority: Major
Reporter: David Hall (Inactive) Assignee: David Hill (Inactive)
Resolution: Fixed Votes: 0
Labels: None
Environment:

Ubuntu 16.04

Issue Links:
PartOf
is part of MCOL-280 Beta issues Closed
Sprint: 1.0.3

 Description   

During the testing of test005 on Ubuntu, DBRMControllerNode crashed for whatever reason. It was automatically restarted, but it left the system in the not ready state. Any query following this event received
ERROR 1815 (HY000) at line 5: Internal error: The system is not yet ready to accept queries
DBRM holds the flag for this state. It is set by DMLProc during startup. When DBRMControllerNode comes back up, it is naturally not ready. Since DMLProc is already running, the flag never gets set to ready.

 Comments   
Comment by David Hall (Inactive) [ 2016-09-02 ]

Added code in ProcMgr to restart DMLProc and to set the SystemQuryReady flag. Also added code to be sure everything's up before moving on.

Changes in processmanager.cpp

Comment by Daniel Lee (Inactive) [ 2016-09-15 ]

Assigned to Mr. Hill for regression test.

Comment by David Hill (Inactive) [ 2016-09-16 ]

In my regression testing on Ubuntu, I didn't have the DBRMControllerNode crash during the test005 run. So cant say the code got test or not for funtionality. But I didnt see this issue in this run, can reopen if it happens in future runs

Generated at Thu Feb 08 02:19:43 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.