slave doesn't set status to STOPPED after max-reconnect failures
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Drizzle |
Fix Committed
|
Undecided
|
Stewart Smith |
Bug Description
Related to bug 743958 (need way to determine master status/information from slave), while testing bug 956200 (slave config file option max-reconnects doesn't work) I found that even after the slave exhausts its attempts to connect to the master, the status is still RUNNING when it should switch to STOPPED and display an error message.
To reproduce:
1. Let slave fail to reconnect max-reconnect times
2. Query the sys_replication tables and see:
drizzle> select * from sys_replication
*******
master_id: 1
status: RUNNING
error_msg:
drizzle> select * from sys_replication
*******
last_applied_
originating_
originating_
Related branches
- David Shrewsbury (community): Disapprove
- Drizzle Trunk: Pending requested
-
Diff: 267 lines (+84/-65)3 files modifiedplugin/slave/queue_producer.cc (+55/-41)
plugin/slave/queue_producer.h (+2/-2)
plugin/slave/replication_slave.cc (+27/-22)
Changed in drizzle: | |
assignee: | nobody → Stewart Smith (stewart) |
status: | Confirmed → Fix Committed |
The linked branch actually fixes this bug by calling setIOState(err_msg, false); if _is_connected is false after all reconnect attempts.