MySQL Forums
Forum List  »  NDB clusters

Re: Wait LCP to ensure durability
Posted by: Mikael Ronström
Date: November 25, 2017 07:33AM

I did some analysis of your problem.
I got a strong feeling that I saw the same problem and fixed
it before.
It took some searching, but I eventually found the following
commit e3f905ccb844db3cf9da7bc6917c49c3fd8d3eeb
Author: Mikael Ronström <mikael.ronstrom@oracle.com>
Date: Fri Jun 2 13:49:31 2017 +0200

WL#10302: BUG#26166901: Possible to get hang of LCP due to a node failure at a very inopportune timing when all nodes waited for this node to complete its last fragments

This bug is fixed in the latest 7.6 DMR, it is currently not fixed
in the latest 7.5 tree.

The problem is that the node failure of node 6 seem to have occurred
at a very inopportune timing. It is a very rare bug.

I noted that you were planning a system restart. I would first attempt
to restart only node 3. This is the node that the master node (nodeid = 5)
is waiting for to report completion.

So if it dies it should clear the wait bits and hopefully the LCP gets
going again. After that you can start up 3 and 6 again.

Options: ReplyQuote


Subject
Views
Written By
Posted
1922
September 28, 2017 03:46AM
823
September 28, 2017 03:50AM
972
September 28, 2017 03:17PM
1593
October 02, 2017 01:56AM
890
October 02, 2017 06:39AM
888
October 04, 2017 03:49AM
1036
November 22, 2017 01:07AM
867
November 22, 2017 01:35AM
839
November 22, 2017 02:23AM
850
November 22, 2017 03:30AM
Re: Wait LCP to ensure durability
832
November 25, 2017 07:33AM
753
November 27, 2017 04:28AM
718
November 27, 2017 08:44AM
763
November 27, 2017 04:16PM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.