Details
-
Bug
-
Resolution: Duplicate
-
Major
-
Cheshire-Cat
-
Untriaged
-
Unknown
Description
Using the current magma-preview repo I configured a two node cluster, added a 256MB magma bucket, and started pillow fight against the bucket. After a bit of time (less than 1/2 hour) errors started occurring:
2020-01-30T13:36:55.689131-08:00 CRITICAL kvstore-13/rev-000000001/localIndex Error occured during memtable flush (Code(10): /Users/stevewatanabe/couchbase/ns_server/data/n_0/data/magmaBucket/magma.1/kvstore-13/rev-000000001/localIndex/sstable.143.data: Unable to open file err=Too many open files). |
Doing some experiments to watch the file count grow I believe it is in the directory:
stevewatanabe @ ~/couchbase/ns_server/data/n_0/data/magmaBucket (magmaFixes) $ ls -lR | wc -l
2715
I did a cbcollect on node0: https://s3.amazonaws.com/cb-engineering/Steve+Watanabe/collectinfo-2020-01-30T213744-n_0%40127.0.0.1.zip
But this issue is very easy to reproduce.
Attachments
Issue Links
- depends on
-
MB-36379 Provide an API to set storage engine file descriptor limit
- Closed