[SERVER-2182] new feature for replica sets : fail over with minority Created: 06/Dec/10  Updated: 16/Nov/21  Resolved: 16/Jun/11

Status: Closed
Project: Core Server
Component/s: Replication
Affects Version/s: None
Fix Version/s: 1.9.1

Type: New Feature Priority: Minor - P4
Reporter: Dwight Merriman Assignee: Kristina Chodorow (Inactive)
Resolution: Done Votes: 3
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Depends
Duplicate
is duplicated by SERVER-2772 Option to "force" reconfiguration on ... Closed
is duplicated by SERVER-1526 Can't add a new server to a replica s... Closed
Participants:

 Description   

if a replica set is partially down, perhaps a majority down, there needs to be an easy way to cut over to the remainder of the set. this can be done today but is administratively tedious.

we could have a command for this.

this requires some research and thinking first; the other part of the set may think it is still up if the network is partitioned. this has to be handled gracefully.



 Comments   
Comment by Kristina Chodorow (Inactive) [ 16/Jun/11 ]

> db.adminCommand(

{replSetReconfig:config, force : true}

)

See http://www.mongodb.org/display/DOCS/Reconfiguring+a+replica+set+when+members+are+down.

Comment by Raviv Pavel [ 16/Jun/11 ]

What's the command?

Comment by auto [ 14/Jun/11 ]

Author:

{u'login': u'kchodorow', u'name': u'Kristina', u'email': u'kristina@10gen.com'}

Message: force reconfig SERVER-2182
Branch: master
https://github.com/mongodb/mongo/commit/aa4edc53fb07a89b74977c9654b0f216b1582b07

Comment by Raviv Pavel [ 25/May/11 ]

Good idea guys, thanks.
Still a fail over with minority would make things much simpler.

Comment by Rasitha Wijesinghe [ 25/May/11 ]

Having an arbiter in a third site is exactly what we are doing. Don't have to worry about manual intervention.

Comment by Dwight Merriman [ 25/May/11 ]

I wonder if the real solution is just to have an arbiter in a third location.

Comment by Raviv Pavel [ 24/May/11 ]

The problem is not only the complexity of the procedure but also the resulting downtime.
There are cases when I know the majority is down because the data center they are in failed.
I should be able to "turn off" the part of the election algorithm that requires a majority.
In most cases this would be a temporary state, until the failing data center come back or until I add more nodes.

Comment by Dwight Merriman [ 06/Dec/10 ]

right. thus the jira.

several howto's here: http://www.mongodb.org/display/DOCS/Replica+Sets

particularly: http://www.mongodb.org/display/DOCS/Reconfiguring+a+replica+set+when+members+are+down

Comment by Arkadiy Kukarkin [ 06/Dec/10 ]

Is there a FAQ entry for manual failover in this case somewhere? The procedure I had to employ was annoying to figure out and required server restarts, which makes me think it may have been less than ideal.

Generated at Thu Feb 08 02:59:12 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.