|
As of SERVER-37270 in 4.1.8, all index builds run in the background. Closing as "Gone away" because there is no longer a need to configure a default.
|
|
It makes sense in the dev environment for the software to create its indexes automatically, so yes this would need to be done in conjunction with a config setting. For production clusters the default could be refuse foreground indexes, but dev clusters could set a config option to allow foreground indexes, which I thought was the idea of this ticket.
|
|
Ernest, it appears that mongomirror was running on a cluster while it was active in production; this situation is not recommended and the behavior of the cluster in that situation is expected. The Epic containing this ticket is to design a new system for index builds that does not exclusively lock a database for their duration – thus removing the problematic behavior of foreground index builds. Once that is in place, the need for a config option will be removed.
|
|
This would be very helpful. I recently was migrating multiple on premise production Mongos into one Atlas cluster using mongomirror. I migrated the first and it was fine. Then I migrated the second, and since it had a foreground index on it, not only was it not usable, it took the previous collections offline too (which we were using in production). This could have been avoided if I could have enforced background index creation.
|
Generated at Thu Feb 08 03:55:48 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.