Open ygoe opened 4 years ago
IIRC:
--format
option to format their (often list-like) outputif you use multiple repos, they have multiple repo directories, so you actually can just use du
on them.
Okay, JSON puts this on a separate line so it can be parsed, too. Can I rely on the JSON format, will it stay like this and not change to a single-line (non-formatted) output?
du
is not supported on Hetzner Storage Boxes via SFTP, so I can't use that. SSH shell access is not available.
No, you can't really rely on that, sorry.
While there are no current plans to change that, I guess we can only guarantee that the json will be valid json, not that it will be formatted in a specific way forever.
Well, could I then suggest adding a general commandline option like --bytes
to always show any size in bytes and not scale them down to smaller numbers?
well, other commands already have --format
so somebody could check whether it can be implemented similar to what we have.
Have you checked borgbackup docs, FAQ, and open Github issues?
Yes
Is this a BUG / ISSUE report or a QUESTION?
Question
System information. For client/server mode post info for both machines.
Your borg version (borg -V).
1.1.11, ready to upgrade if necessary
Operating system (distribution) and version.
Ubuntu 18.04, 20.04
Hardware / network configuration, and filesystems used.
SSH remote
How much data is handled by borg?
Gigabytes
Full borg commandline that lead to the problem (leave away excludes and passwords)
borg info $REPOSITORY
Describe the problem you're observing.
The output looks like this:
I want to parse the last number and use it for my monitoring. Multiple borg backups go to the same remote SSH account so I cannot simply fetch the disk usage from there, it's the sum of all repositories. To know which of the repositories is large, I need separate numbers.
My borg wrapper is a Bash script. I could probably parse those human-friendly formatted numbers but I'd like to know if there's a commandline switch to spare me and borg the additional work. Haven't tried the JSON format but I'd like to skip the dependency on
jq
if possible, and this is a simple case for parsing with grep and awk.The documentation doesn't list such a parameter. I wanted to inspect the source code but could not find the relevant part for the info command.