[pgpool-general: 5327] Re: URGENT: Split-brain scenario...how to prevent?

Lazaro Garcia lazaro3487 at gmail.com
Sat Feb 11 01:32:16 JST 2017


Is fail_over_on_backend_error on or off?

 

Try with off?

 

Regards.

 

De: pgpool-general-bounces at pgpool.net
[mailto:pgpool-general-bounces at pgpool.net] En nombre de David Sisk -X (dsisk
- TEKSYSTEMS INC at Cisco)
Enviado el: viernes, 10 de febrero de 2017 05:00 p. m.
Para: pgpool-general at pgpool.net
Asunto: [pgpool-general: 5325] Re: URGENT: Split-brain scenario...how to
prevent?
Importancia: Alta

 

Can anyone provide any insight into how to prevent this split brain scenario
from happening?  Any responses appreciated.

 

Thanks and regards,

Dave Sisk

 


 


 

 

From: pgpool-general-bounces at pgpool.net
<mailto:pgpool-general-bounces at pgpool.net>
[mailto:pgpool-general-bounces at pgpool.net] On Behalf Of David Sisk -X (dsisk
- TEKSYSTEMS INC at Cisco)
Sent: Thursday, February 9, 2017 2:38 PM
To: pgpool-general at pgpool.net <mailto:pgpool-general at pgpool.net> 
Subject: [pgpool-general: 5321] URGENT: Split-brain scenario...how to
prevent?

 

Hi folks.PGPool 3.5.4 with Postgres 9.3, streaming replication mode with
primary and one standby.  I found a scenario that caused a split-brain
(luckily, it's in a lab environment instead of a prod environment).

 

1)      Auto-failover occurred, original standby is now primary.

2)      Prior primary/eminent standby at status 3 NOT reset/sync'd in any
way yet (not replicating and not in standby mode).

3)      pcp_attach_node 0 not only attaches the faulty standby, it actually
promotes the faulty standby back to primary! :-0

 

What configuration parameters will prevent this from happening?  I'd prefer
to get an error from pcp_attach_node and/or have the node stay at status 3.

 

Here are the health checks I have defined:

health_check_period = 10# NON-DEFAULT

health_check_timeout = 20

health_check_user = 'postgres'  # NON-DEFAULT

health_check_password = 'postgres'  # NON-DEFAULT

health_check_database = ''

health_check_max_retries = 2# NON-DEFAULT

health_check_retry_delay = 10# NON-DEFAULT

 

*  Not sure if this is used for load-balancing only, or also used to manage
node status

sr_check_period = 10

sr_check_user = 'postgres'  # NON-DEFAULT

sr_check_password = 'postgres'  # NON-DEFAULT

sr_check_database = 'postgres'

delay_threshold = 100000# NON-DEFAULT

 

 

 

 





 

		

David Sisk

Engineer - Software

 <mailto:dsisk at cisco.com> dsisk at cisco.com

Tel: 

Cisco Systems, Inc.

7025-6 Kit Creek Road PO Box 14987
RESEARCH TRIANGLE PARK
27709-4987
United States
cisco.com

	

 


Think before you print.


This email may contain confidential and privileged material for the sole use
of the intended recipient. Any review, use, distribution or disclosure by
others is strictly prohibited. If you are not the intended recipient (or
authorized to receive for the recipient), please contact the sender by reply
email and delete all copies of this message.

Please  <http://www.cisco.com/web/about/doing_business/legal/cri/index.html>
click here for Company Registration Information.

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20170210/4fe89997/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 42730 bytes
Desc: not available
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20170210/4fe89997/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.gif
Type: image/gif
Size: 134 bytes
Desc: not available
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20170210/4fe89997/attachment-0001.gif>


More information about the pgpool-general mailing list