Inconsistent handling of "Deep" storage schema?

[zmeggyesi]

I've been bugged by a persistent problem for a few days now, ever since I added an external drive to the Raspi running ZM: events would be created, JPGs saved, videos recorded, then a few minutes later, they would disappear from the drive without a trace, while remaining in the database.

For instance, just after running Record for a few minutes, I have this on the external drive:

Code: Select all

root@observer:/mnt/external# ll 6/18/10/04/06/36/31/ | wc -l
996

which all goes away after zmaudit runs. Since I have level 9 debugging on zmaudit, I can see in the log that it locates the appropriate directories using the globs, but then for whatever reason decides that the event ID symlink is not valid (it is), and deletes everything.

Even more strange, the issue doesn't occur if I set the storage schema to "Shallow" - auditing runs and the events survive just as intended.

[knight-of-ni]

Is your eternal drive formatted with a native Linux filesystem that supports symbolic links, such as ext4?

[zmeggyesi]

Hi,

Yes, I'm using Ext4 on the external drive.

When executing ls -lah (which ll is an alias for, just to be clear), I can see the link pointing to the right directory.

[iconnor]

It's probably the bug we had that deletes events.

First off upgrade to 1.32.1
Second turn off zmaudit and run it manually or via cron or whatever (because it isn't required and is a resource hog)

If that doesn't fix it, check your filters. Often will will do greater than -7days... which means anything newer instead of anything older.