Bug #11747 Floating Point Execption was caught by data node after "5 restart" issued
Submitted: 5 Jul 2005 22:45 Modified: 2 Sep 2005 17:25
Reporter: Jonathan Miller
Status: Closed
Category:Server: Cluster Severity:S2 (Serious)
Version:5.1.0-wl2325-wl1354-new OS:Linux (Linux)
Assigned to: Tomas Ulin Target Version:

[5 Jul 2005 22:45] Jonathan Miller 
Description:
I was wanting to move the Sabre test off to ndb10 allowing ndb08 and ndb09 to do nothing
more then processing the test request. To do this I needed to alter the config.ini to add
ndb10 as an API NODE.

After altering the config.ini, I killed and restarted ndb_mgmd to load the new
config.ini. I then started to cycle the data nodes one by one. When I restarted the last
data node "node 5" it recieved a signal 8 "SIGFPE".

Date/Time: Tuesday 5 July 2005 - 22:32:02
Type of error: error
Message: No message slogan found
Fault ID: 0
Problem data: Signal 8 received
Object of reference: main.cpp
ProgramName: /home/ndbdev/jmiller/builds/libexec/ndbd
ProcessID: 10707
TraceFile: /space/run/ndb_5_trace.log.4
Version 5.1.0 (a_drop5p3)
***EOM***

Error log and trace file have been saved to ndb09:/space/bug#### where #### is = to the
number of this report.

No core files assosiated with this failure.

How to repeat:
Do not have a repetable test case.
[2 Sep 2005 17:25] Jonathan Miller 
Unable to reproduce. Tried with latest build on Alcatel clone.