| Issue 7: | move_role sets server to slave of self if mmmd_agent isn't running |
1 of 22
Next ›
|
| 2 people starred this issue and may be notified of changes. | Back to list |
Sign in to add a comment
|
What steps will reproduce the problem? 1. Kill mmmd_agent on the reader of the monitored servers. 2. switch the writer role to that server. 3. switch the writer role back away from that server. What is the expected output? What do you see instead? I think if the agent on the server can't be contacted, MMM should complain and refuse to go ahead. Instead it moves the writer role, but when it tries to move it back, it complains (though it still seems to move it back, actually) and then somehow the server ends up being set as a slave of itself. Then the server goes into REPLICATION_FAIL state. |
||||||||||
,
Nov 01, 2008
I can't reproduce the problem: db1(10.20.30.31): master/ONLINE. Roles: reader(10.20.30.52;), writer(10.20.30.50;) db2(10.20.30.32): master/ONLINE. Roles: reader(10.20.30.53;) db3(10.20.30.33): slave/ONLINE. Roles: reader(10.20.30.51;) on db2: kill `cat $MMMD_AGENT_PIDFILE` on mon: mmm_control move_role writer db2 Command sent to monitoring host. Result: ERROR: Can't reach agent daemon on 'db2'! Can't move roles there! |
|||||||||||
,
Jan 14, 2009
(No comment was entered for this change.)
Labels: Release-1.0
|
|||||||||||
 Sign in to add a comment
|
|||||||||||
