-
Type:
Investigation
-
Resolution: Unresolved
-
Priority:
Major - P3
-
Affects Version/s: 4.2.15
-
Component/s: mongorestore
-
None
-
1,013
Problem Statement/Rationale
What is going wrong? What action would you like the Engineering team to take?
mongorestore of oplog from archive hangs without providing any indication of what is happening or what is wrong.
Steps to Reproduce
How could an engineer replicate the issue you're reporting?
Create an archive from an oplog:
mongodump --host <node_02_hostname> --port <port> --db local --collection oplog.rs --archive=oplog_dump.02.archive
Now, try to restore into a different mongod instance.
This just hangs:
> mongorestore --archive=oplog_dump.02.archive --host=localhost --port=27017 2021-11-07T15:46:38.556+0200 preparing collections to restore from
Now with verbose – also hangs:
> mongorestore --archive=oplog_dump.02.archive --host=localhost --port=27017 -v
2021-11-07T15:52:00.901+0200 using write concern: &{majority false 0}
2021-11-07T15:52:01.010+0200 archive prelude local.oplog.rs
2021-11-07T15:52:01.012+0200 preparing collections to restore from
2021-11-07T15:52:01.012+0200 found collection local.oplog.rs bson to restore to local.oplog.rs
2021-11-07T15:52:01.012+0200 found collection metadata from local.oplog.rs to restore to local.oplog.rs
After investigation, it seems that this was trying to create collection local.oplog.rs, which hung. (I tried this with restore target = single instance, and = replicaset.)
Now with nsFrom+nsTo – now we get an error:
> mongorestore --archive=oplog_dump.02.archive --host=localhost --port=27017 --nsFrom="local.oplog.rs" --nsTo "test.test" 2021-11-07T15:52:40.355+0200 preparing collections to restore from 2021-11-07T15:52:40.376+0200 reading metadata for test.test from archive 'oplog_dump.02.archive' 2021-11-07T15:52:40.445+0200 finished restoring test.test (0 documents, 0 failures) 2021-11-07T15:52:40.445+0200 Failed: test.test: error creating collection test.test: error running create command: (Location50001) autoIndexId:false is not allowed for collection test.test because it can be replicated 2021-11-07T15:52:40.445+0200 0 document(s) restored successfully. 0 document(s) failed to restore.
It looks like this was always trying to create the new collection with autoIndexId:false – which is not allowed for a collection that can be replicated – because the dumped collection was the oplog.
Pre-creating the collection (without autoIndexId:false) solved this.
Expected Results
What do you expect to happen?
Instead of hanging, this should have failed with a meaningful error message.
Actual Results
What do you observe is happening?
Additional Notes
Any additional information that may be useful to include.