bupstash.io will update the server and people won't notice
this was kind of naive and not how the real world works :D
for future versions and 1.0 my plan is to essentially never break backwards compatibility
and just try hard in the server to keep supporting all old clients
my future compatibility goals are the user should be able to use whatever version they want
and the server just deals with it, this is the reality because I want bupstash to be distributed by linux distros
and have that version work with the managed repositories
at that point we lose a lot of control of the versions
if needed
Hi folks, just tried bupstash out and I'm really liking it. I'm backing up my personal and work laptop to my NAS now and it's working great.
I've had the following issue happen twice so far, wondering if anybody got it as well?
Connection to <host> closed by remote host.
bupstash put: remote disconnected
Its curious it isn't printing any more information - normally when the server side exits it prints a message
@tomleb:matrix.org which version of bupstash is running on the server and client?
Unfortunately I don't have much more information. I can see what the last file was being uploaded, a small jpg image. For some reason it's outputed multiple times instead of only once.
export BUPSTASH_REPOSITORY_COMMAND='ssh sh -c "you@server bupstash server /path/to/repo 2 >> /tmp/bupstash-errors.log"'
if that makes sense
then if the error happens whatever was printed to stderr will be on your nas in that file
assuming it is printing anything
sorry, 'bupstash serve' not 'bupstash server'
Alright, I'll try that. I forgot to take note of the time these issues happened. I've got SOME sshd logs but I can't correlate with time :/ Let me try again.
Hopefully
export BUPSTASH_REPOSITORY_COMMAND='ssh <host> sh -c "bupstash serve /path/to/repo >& /tmp/bupstash-errors.log"' works.. The default shell is csh so apparently I need to use >& to redirect. Hopefully that works lol.
Gave it 2 tries and wasn't able to reproduce, will try again tomorrow.
that*
It doesn't look like it so I'll do more tests tomorrow, switch to
bash instead as well
I happened again. Here's the error message from sshd logs:
Jan 29 18:54:08 bupstash sshd[60127]: Fssh_ssh_dispatch_run_fatal: Connection from user root 192.168.1.38 port 33420: message authentication code incorrect
The redirection didn't see to work so I don't have more logs than this. I'll try again later.
i.e., it looks like ssh is the one disconnecting your backup
since it detects a modified packet and disconnects
quite strange - I wonder what the cause could be
Yeah, well it's probably not a bupstash issue then. Will see if I can figure out what's wrong, thanks!
I'm doing the backup over wifi and I can see "wlp3s0: WPA: Group rekeying completed " in my journald logs, so I'll investigate my AP
How are you folks using tags with bupstash? Any funky usage of them? I've added a kernel and os tag to keep track of the kernel version and os (eg: fedora 37) of a particular backup. Are there other good uses of them?
1 reply
I also include the nodes ssh-public key as a unique ID for the case of hostname collisions
Oh and a url like id for the script/program/interaction that created the backup so that I can easier track down the source of the backup
1 reply
may be there could be a place to share tips in a more persistent fashion...