queryLog.py out of control with big log files

[Marin BERNARD]

Hi,

In an attempt to increase on-disk log file retention, we adjusted the log file size parameter to 100 000 000 bytes. Since then, trying to browse records from any log from the GUI results in the /usr/local/opnsense/scripts/systemhealth/queryLog.py process eating all available memory (several GB) until complete RAM + swap exhaustion, making the box unresponsive until restarted.

Should a bug be opened for this on GitHub ? IMHO, this is a serious issue.

Thanks,

[franco]

Note that scan size is more or less linear with clog file size. With big clog files this will happen and cannot be fixed as we aways have to scan the full buffer (if it is filling up).

20.7 will be able to switch to log archive rotation.


Cheers,
Franco

[Marin BERNARD]

This would be great!
 Many thanks for taking the time to answer.