Issue 1119 in tungsten-replicator: auto-recovery stops after only one try

14 views
Skip to first unread message

tungsten-...@googlecode.com

unread,
Aug 14, 2015, 7:20:14 PM8/14/15
to tungsten-repl...@googlegroups.com
Status: New
Owner: ----
Labels: Type-Defect Priority-Medium FoundIn

New issue 1119 by Dominiqu...@unige.ch: auto-recovery stops after only one
try
https://code.google.com/p/tungsten-replicator/issues/detail?id=1119

What steps will reproduce the problem?

1. configure auto_recovery parameters eg:

"repl_auto_recovery_delay_interval": "20s",
"repl_auto_recovery_max_attempts": "3",
"repl_auto_recovery_reset_interval": "80s"

2. stop the database (mysql)

3. wait until after two times auto-recovery-delay-interval

4. start the database


What is the expected output?

The replicator goes offline when mysql is stopped,
then tries twice to reconnect, and on the third try it goes back online
because mysql is up again.

What do you see instead?

The replicator tries to reconnect only once. Then
only a manual trepctl online will put the replicator back online.

This is the behaviour described in issue 784
https://code.google.com/p/tungsten-replicator/issues/detail?id=784
and, in a less obvious way, in the manual
http://docs.continuent.com/tungsten-replicator-4.0/operations-autorecovery.html

But it does not make much sense:
If a manual "treplctl online" works why should not an automatic reconnection
be done as well?
The argument of issue 784 applies here as well:
"A lot of users write watchdog scripts to handle these, which seems like a
waste of time."

What version of the product are you using?

./tools/tpm query version
4.0.0-18


mysql -V
mysql Ver 14.14 Distrib 5.5.44, for debian-linux-gnu (x86_64) using
readline 6.3

On what operating system?

head -1 /etc/os-release
PRETTY_NAME="Debian GNU/Linux 8 (jessie)"


Please provide any additional information below.

Notes:

See a full description, explanation and remarks in Tungsten Replicator
Discuss:
https://groups.google.com/d/msg/tungsten-replicator-discuss/7mHu0sXarto/XeHS5wN3AgAJ



--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
Reply all
Reply to author
Forward
0 new messages