Hi!
I’m running Bareos for backing up my machines.
Before starting the backup, Bareos runs a influx backup $backupdir
command and cleans the backup directory afterwards.
This was nice until the influxdb grew bigger and I ran out of disk space in the backup directory.
Now that Bareos supports virtual files that are fed from the output of, for example, an external command, I’d like to write a Bareos plugin that makes use of this.
Therefore I need a way to stream the influx backup to stdout. Since influx backup
does not support this, I thought of something like
for org in $(influx org list --json | jq -r .[].name); do
for bucket in $(influx bucket list -o "${org}" --json | jq -r '.[].id'); do
influxd inspect export-lp --bucket-id "${bucket}" \
--engine-path /var/lib/influxdb/engine --output-path -
done
influx export all -o "${org}"
done
sqlite3 /var/lib/influxdb/influxd.sqlite .dump
(this example is not meant to work, just to get the idea)
and additionally saving /var/lib/influxdb/influxd.bolt
.
Is there something missing in order to be able to restore all buckets, tasks, orgs, users, notebooks, dashboards, graphs, etc.?
What’s the right restore procedure? influx write --bucket $mybucket
, restoring the sqlite3/bolt databases, what else?
Thanks in advance!