search

canonical / mongodb-operator

Charmed solution for MongoDB
Apache License 2.0
10 stars 13 forks source link

Import of big data sets #398

Open dmitry-ratushnyy opened 2 months ago

dmitry-ratushnyy commented 2 months ago

Steps to reproduce

  1. Deploy MongoDB cluster, for example on LXD
  2. Obtain connection URI
  3. Install mongosh on host machine in case of LXD or copy your data set into juju unit
  4. Using "charmed-mongodb.mongoimport" (from inside of juju unit) or mongoimport (from host machine) start import data
  5. Periodically check free space on disk.

Expected behavior

Import of 70 GB of data should not consume 500+GB of disk space

Actual behavior

Disk space is being consumed very quickly. This most likely happens due to audit logs. Log rotation and allowing users to opt out from audit logging should solve the problem

Versions

stable/164

Operating system:

Juju CLI:

Juju agent:

Charm revision:

LXD:

Log output

Juju debug log:

Additional context

github-actions[bot] commented 2 months ago

https://warthogs.atlassian.net/browse/DPE-4058

dmitry-ratushnyy commented 2 months ago

We are working on resolving this issue