Significant increase on the database size (4x the size)

[b10c77]

I'm running a few nodes and noticed that from time to time, the database size goes out of control and grows until it fills the disk.

Right now the size is ~2.1 GB on 2 of my nodes and 8.2 GB on another one. I'm including some screenshots of the sqlite directory which seem to be where the significant size increase happens.

"Normal" size (~2.1 GB)

Abnormal size (~8.2 GB)

[gregorycollins]

a) what version are you running? b) When you kill chainweb-node, what signal are you sending? The binary likes to be killed with SIGINT -- if you kill with SIGKILL it won't clean up its WAL files.

[larskuhtz]

@smartins I looks like the increased storage is due to sqlite WAL file. By default sqlite doesn't release or prune those files during operation or at startup. Did you try to use SIGINT when stopping or restarting nodes? That fixed the issue for us when we tried to reproduce it.

[larskuhtz]

If you use systemd you may use the following line in the [Service] section:

KillSignal=SIGINT