Juan Calderon-Perez
Juan Calderon-Perez
@hengfeiyang @prabhatsharma I did some debugging today, when ZO gets restarted/upgraded and a query is executed the error always happens on any interval after 3h, so for 6h, 8h, 12h,...
@hengfeiyang Unfortunately I don't have access to the logs at home.
@hengfeiyang One thing I noticed is that ZO is using an older version of arrow-rs v36.0.0. The latest version is v39.0.0 and there has been a lot of bug fixes...
@prabhatsharma That sounds accurate. The query tries to get ~4 files, and there is 4 files in the folder but the names are slightly different than the ones ZO is...
@hengfeiyang This can be closed if you want. I have no way of sharing the logs. I'm also unable to replicate this issue using a local ZO instance. All I...
@hengfeiyang After checking yesterday... Turns out it's not a network attached storage. The data is sitting in a mount managed by lvm/devicemapper, it also had `xfs,noatime` as mount options. Removing...
Just an update, I was using the ZO_DATA_DIR. Removing that seems to have fixed the problem. Will confirm tomorrow after rebooting the container with +1day of data.
Nevermind, it still happens. I cant query any data older than +2hr. The files OpenObserve is looking for don't match the ones on disk.
I'm running v0.4.5, the bug is not the no data one. I get a "Error parquet file not found" since like 0.3.x Basically every time I upgrade the container, any...
@hengfeiyang I have no way of accesing this instance from outsite of work. I have been trying to replicate the issue at home. No luck so far replicating it.