Is there a way to tell which files and directories contributed most to a new snapshot? #1478
-
|
I take fairly frequent snapshots, and I find the "Added to the repo" result to often be far larger than I expect, given my backup sources and system activity, but i haven't been able to find any way to track down why, no way to get rustic to tell me, say, the top 10 largest files/directories in this backup run. Even the interactive snapshot viewer is of no use here, because directories always report 0 size. I suspect some obscure cache that isnt in XDG_CACHE_HOME and doesn't use CACHEDIR.TAG is causing frequent changes, but short of i dont know writing a script to manually go through and examine all the snapshot data and compile the sizes myself, i dont think theres an easy way to find out? And I dont want to write such a script, it feels like rustic already knows this information and already keeps a running total of size and new files/directories and changed ones, and I don't want to have to manually re-do it after every backup run. |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 6 replies
-
|
@DianaNites Yes, in principal rustic has all information to show differences in such a way - it is simply not implemented yet. I thought about an interactive About the interactive |
Beta Was this translation helpful? Give feedback.
-
|
@DianaNites in #1480 I made a WIP-implementation of an interactive diff. This already works pretty well for me and the option to compute added/removed sizes recursively should help you with the topic of this discussion. Can you test that PR and give me feedback? |
Beta Was this translation helpful? Give feedback.
-
|
The interactive diff command (which includes changes to the repository and is able to calculate summaries over directories) has now been added to master, see #1480. |
Beta Was this translation helpful? Give feedback.
The interactive diff command (which includes changes to the repository and is able to calculate summaries over directories) has now been added to master, see #1480.