Question about version for borg over SSH #8621
-
|
Hello, I hesitate in 2 senarios to backup in other machine my data with BORG. In the second case, i need borg on the server. I try to find a way for the long time :) |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 3 replies
-
|
Using a ssh:// repo should be faster usually, because in that case, there will be a "borg serve" repo-side process that can access the repository locally without having to transfer the data over the network (e.g. for "borg check" [repository part]). borg version does not need to be precisely the same on client and server, but they need to be compatible. borg2 is a breaking release and not compatible with borg1 (except that it supports transferring data from borg1 repos). For now, you should use borg 1.x for production. If you like, you can set up an additional backup using borg2 to help us testing. |
Beta Was this translation helpful? Give feedback.
-
|
ok, noted for the borg check. I noted borg2 not compatible with borg1 but i want to use borg for a long time and analyze if i have not the same version between client and the server which stores the backups (when borg2 will be released) |
Beta Was this translation helpful? Give feedback.
-
|
borg create is usually rather quick in general, but of course it depends on how much new/changed data you generate between backups. but borg check always has to check everything, so it is rather time consuming in general. |
Beta Was this translation helpful? Give feedback.
-
|
For Fedora and Fedora EPEL, my goal is to provide borg1 + borg2 in parallel so you can install either version (but probably not both). Not sure how realistic this plan is given my time constraints but that is my preferred outcome. |
Beta Was this translation helpful? Give feedback.
-
|
Oh it's a good news Felix ! Thanks for this answer for the Red Hat branch lovers :) |
Beta Was this translation helpful? Give feedback.
-
|
@FelixSchwarz maybe look at how the debian packages work, IIRC you can have them both installed. |
Beta Was this translation helpful? Give feedback.
Using a ssh:// repo should be faster usually, because in that case, there will be a "borg serve" repo-side process that can access the repository locally without having to transfer the data over the network (e.g. for "borg check" [repository part]).
borg version does not need to be precisely the same on client and server, but they need to be compatible. borg2 is a breaking release and not compatible with borg1 (except that it supports transferring data from borg1 repos).
For now, you should use borg 1.x for production.
borg2 is still in beta and can break between beta releases.
If you like, you can set up an additional backup using borg2 to help us testing.