Open takayuki-nagai opened 1 year ago
We now support RHEL 9 https://apache.jfrog.io/ui/repos/tree/General/couchdb-rpm/el9 (as Almalinux 9) and we just released 3.3.3. Give that a try, wonder if it will show the same memory usage pattern you saw before.
Im having similar issues, running Hyperledger Fabric inside k8s on Fedora release 35 (Thirty Five)
.
Memory sky rockets to multiple GBs, then gets OOM killed.
Zero traffic is send to service.
Also tried latest
docker tag and 3.3.3
Description
I've noticed that CouchDB's memory usage increases abnormally when Hyperledger Fabric(HLF) is running on certain operating systems.
I deployed HLF v2.5.1, which uses CouchDB v3.3.2, on a VM with RHEL9 installed. (In HLF, CouchDB is provided as a docker image; base OS is debian gnu/linux 11.) Then the memory usage of the CouchDB container reached around 1.5GB. I also found that the same phenomenon occurs when deploying HLF v2.4, which uses CouchDB v3.1.1, on RHEL9. On the other hand, when deploying HLF on RHEL8 or Ubuntu, CouchDB memory usage was only about 50MB.
I know couchdb doesn't support RHEL9 yet. Will there be any possibilities that deploying a docker container including couchdb on RHEL9 causes high memry usage?
Your Environment