Bug #43853 Detected GCP stop is just a INFO but should be a WARNING at least

• Submitted: 25 Mar 2009 8:40
• Modified: 24 Jun 2009 9:46
• Reporter: Oli Sennhauser
• Status: Closed
• Category: MySQL Cluster: Cluster (NDB) storage engine
• Severity: S3 (Non-critical)
• Version: mysql-5.1-telco-7.0
• OS: Any
• Assigned to: jack andrews
• CPU Architecture: Any

[25 Mar 2009 8:40] Oli Sennhauser

Description:
A GCP stop detection is actually written to the cluster log as an INFO:

2009-03-23 03:13:45 [MgmSrvr] INFO     -- Node 4: Detected GCP stop(1)...sending kill to [SignalCounter: m_count=1 0000000000000008]

GCP stop usually can cause serious troubles to the cluster and IMHO should be noted as a WARNING at least.

How to repeat:
Provoke a GCP stop.

Suggested fix:
Log level WARNING or higher.

[29 Apr 2009 10:00] jack andrews

hi mganus, how's this patch?

$ quilt diff
Index: 70/storage/ndb/src/kernel/blocks/dbdih/DbdihMain.cpp
===================================================================
--- 70.orig/storage/ndb/src/kernel/blocks/dbdih/DbdihMain.cpp   2009-04-29 19:59
:00.026000000 +1000
+++ 70/storage/ndb/src/kernel/blocks/dbdih/DbdihMain.cpp        2009-04-29 19:59
:04.463000000 +1000
@@ -13192,7 +13192,7 @@
       signal->theData[0] = 2305;
       sendSignal(rg, GSN_DUMP_STATE_ORD, signal, 1, JBB);

-      infoEvent("Detected GCP stop(%d)...sending kill to %s",
+      warningEvent("Detected GCP stop(%d)...sending kill to %s",
                 m_gcp_save.m_master.m_state, c_GCP_SAVEREQ_Counter.getText());
       ndbout_c("Detected GCP stop(%d)...sending kill to %s",
                m_gcp_save.m_master.m_state, c_GCP_SAVEREQ_Counter.getText());
@@ -13204,7 +13204,7 @@
       /**
        * We're waiting for a COPY_GCICONF
        */
-      infoEvent("Detected GCP stop(%d)...sending kill to %s",
+      warningEvent("Detected GCP stop(%d)...sending kill to %s",
                 m_gcp_save.m_master.m_state, c_COPY_GCIREQ_Counter.getText());
       ndbout_c("Detected GCP stop(%d)...sending kill to %s",
                m_gcp_save.m_master.m_state, c_COPY_GCIREQ_Counter.getText());
@@ -13254,7 +13254,7 @@
     /**
      * We're waiting for a GCP PREPARE CONF
      */
-      infoEvent("Detected GCP stop(%d)...sending kill to %s",
+      warningEvent("Detected GCP stop(%d)...sending kill to %s",
                 m_micro_gcp.m_state, c_GCP_PREPARE_Counter.getText());
       ndbout_c("Detected GCP stop(%d)...sending kill to %s",
                m_micro_gcp.m_state, c_GCP_PREPARE_Counter.getText());
@@ -13281,7 +13281,7 @@
     }
     case MicroGcp::M_GCP_COMMIT:
     {
-      infoEvent("Detected GCP stop(%d)...sending kill to %s",
+      warningEvent("Detected GCP stop(%d)...sending kill to %s",
                 m_micro_gcp.m_state, c_GCP_COMMIT_Counter.getText());
       ndbout_c("Detected GCP stop(%d)...sending kill to %s",
                m_micro_gcp.m_state, c_GCP_COMMIT_Counter.getText());

[29 Apr 2009 11:50] Magnus Blåudd

That should solve the problem.

[12 Jun 2009 10:27] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/76172

2937 jack andrews	2009-06-12
      Bug #43853  	Detected GCP stop is just a INFO but should be a WARNING at least
        . just changed infoEvent to warningEvent in 4 places

[15 Jun 2009 9:51] Bugs System

Pushed into 5.1.35-ndb-7.0.7 (revid:jack@asus-20090615083619-6n3n3a98gtd1xjx9) (version source revid:[e-mail]-20090612102517-0255ptalu54yrer7) (merge vers: 5.1.34-ndb-7.0.7) (pib:6)

[24 Jun 2009 9:46] Jon Stephens

Documented bugfix in the NDB-7.0.7 changelog as follows:

        Previously, a GCP STOP event was written to the cluster log as
        an INFO event. Now it is logged as a WARNING event instead.