[SERVER-427] Replication log directory should also be config file or user/cli configurable. Created: 17/Nov/09  Updated: 02/Aug/18  Resolved: 17/Nov/09

Status: Closed
Project: Core Server
Component/s: Admin, Usability
Affects Version/s: 1.1.3
Fix Version/s: None

Type: Task Priority: Major - P3
Reporter: Alexey Zilber Assignee: Eliot Horowitz (Inactive)
Resolution: Duplicate Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Linux


Participants:

 Description   

Same reasons as the previous. Currenty, with MySQL and EBS volumes for snapshot servers we have the relay log directory on a local ephemeral volume on slaves. If the slave falls behind (which it will during a snapshot) the relay log directory will start filling up. It would be bad for that data to hit the ebs volume as that would kill the slave. I'm assuming at least some of these issues would also affect MongoDB, considering AWS EBS performance can be flakey. This issue might also affect iSCSI volumes as well if there's a lot of load.

-Alex



 Comments   
Comment by Eliot Horowitz (Inactive) [ 17/Nov/09 ]

the replication logs are just in the local database.
so you can put that local database wherever you want
i.e.
/data/db/
local -> /mnt/other/

Comment by Alexey Zilber [ 17/Nov/09 ]

SERVER-308 is a good idea, but the replication logs should still be in a different location because that doesn't solve the EBS problem, or for volumes with limited bandwidth (iSCSI) vs local/attached storage.

Comment by Eliot Horowitz (Inactive) [ 17/Nov/09 ]

I think SERVER-308 is a better solution for this.
Solves a larger general problem, and for this you can just link local anywhere you want.

Generated at Thu Feb 08 02:54:04 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.