[MCOL-3414] Failover code for dbroot move fails with storagemanager Created: 2019-08-05  Updated: 2020-12-17  Resolved: 2020-12-17

Status: Closed
Project: MariaDB ColumnStore
Component/s: ?, Storage Manager
Affects Version/s: Icebox
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Ben Thompson (Inactive) Assignee: Ben Thompson (Inactive)
Resolution: Won't Fix Votes: 0
Labels: None

Issue Links:
PartOf
is part of MCOL-3611 Make sure oam does the right thing on... Closed
Epic Link: ObjectStore support

 Description   

The current failover code when a pm module goes down will try to move a dbroot to another node in a variety of cases. this code returns on internal storage. however the current checks do not verify proper behavior for StorageManager storage type.

from bool Oam::autoMovePmDbroot(std::string residePM) :

    writeLog("autoMovePmDbroot: " + residePM, LOG_TYPE_DEBUG );
 
    string DBRootStorageType;
 
    try
    {
        getSystemConfig("DBRootStorageType", DBRootStorageType);
    }
    catch (...) {}
 
    string DataRedundancyConfig = "n";
 
    try
    {
        getSystemConfig( "DataRedundancyConfig", DataRedundancyConfig);
    }
    catch (...)
    {
        DataRedundancyConfig = "n";
    }
 
    if (DBRootStorageType == "internal" && DataRedundancyConfig == "n")
        return 1;



 Comments   
Comment by Patrick LeBlanc (Inactive) [ 2019-08-15 ]

The weird stuff OAM does around failover might work now with the Ownership class written, and if the cluster is set up a particular way. Need to think about this one.

Comment by Ben Thompson (Inactive) [ 2020-12-17 ]

OAM removed from product.

Generated at Thu Feb 08 02:42:33 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.