[SERVER-22221] Question about sharded cluster configuration Created: 19/Jan/16  Updated: 21/Jan/16  Resolved: 21/Jan/16

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

Type: Question Priority: Major - P3
Reporter: Andras Toth Assignee: Ramon Fernandez Marina
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Participants:

 Description   

Hello there,
We have a question about our newly started project, Shardimage, which is more complex than the regular ones. Therefore here comes a little bit of explanation about our plans.
We would like to create 2 clusters on two continents, with the help of MongoDB v3.2. This system would be built by:

  • 5 config servers (from these servers, we would like to put 3 config servers in the USA and 2 in Germany)
  • 4 query routers (2-2 query servers both in the USA and in Germany.)
  • 12 servers used by our service

From these 12 Shardimage servers we would like to create 4 ReplSet replication databases. These dbs would be built by 3-3 Shardimage servers (2 in the US, one in Germany).

The question is: in your opinion, do you think this is feasible? And if yes, what will happen if the connection is lost between the Europian and US server park?
Will the system be working correctly? Both system ReplSets will have a master?

If the reply is okay to both of the question below and the connection is restored between the server parks, then will the US and German master be selected automatically? And will the inconsistent data be updated by automatically or should it be updated by manually?



 Comments   
Comment by Ramon Fernandez Marina [ 21/Jan/16 ]

t.andr, the setup you describe is a very reasonable one. Some things you need to keep in mind:

  • In the event of a partition between the US and EU data centers the US side of the cluster will continue to operate, but the EU side will be unavailable.
  • Once connectivity is restored, the EU side will re-join the cluster automatically.
  • If a European node has fallen too far behind you'll need to manually resync that node. If you anticipate long service disruptions between continents you should plan accordingly by configuring a larger oplog from the beginning.

If what you need is for the US and EU parts of the cluster to continue to operate in the event of a partition you can look at tag aware sharding to see if it meets your needs.

Please note that the SERVER project is for reporting bugs or feature suggestions for the MongoDB server. For MongoDB-related support discussion please post on the mongodb-user group or Stack Overflow with the mongodb tag, where your question will reach a larger audience.

A question like this involving more discussion would be best posted on the mongodb-user group. Please be sure to include details about your use case and your specific needs, and what kind of availability you require in the event of trans-Atlantic network partitioning.

See also our Technical Support page for additional support resources.

Regards,
Ramón.

Generated at Thu Feb 08 03:59:46 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.