Where Are Hashes Stored on The Client?

Here’s my goal. I would like to be able to blow-away an installation of a client, re-install the software client and still retain some sort of “memory” of what was backed up in terms of blocks on the hard disk. If this is not possible, please let me know.

Does UrB store any of the already uploaded hashes or markings of what blocks were baked up in some sort of file form? If so, where are these stored so I can retrieve them from a previous backup and just replace or add them to the client installation or is this not how it works for CBT?

When you re-install a client from scratch, it seems to check the entire drive all over again, like the full image backup was not done.

Thoughts? Thanks.

For image backup, changes are stored in CBT (without it, the drive is fully read each time).

For file backup, changes are stored in the file system journaling (without it, all files are read each time).

Ah, I understand your motivation for the request above, but the good news is that while the entire disk/all files would be read, only changes are transferred to the server (for file backup use block differences transfer mode, for image backup it’s automatic).

For file backup, you would see a message such as this in the logs:

123 GB of files were already present on the server and did not need to be transferred

While I’m here …

How can i verify that “file journaling” is “ON” or that it is being used.

How can I verify that I have the “CBT” version of the client?

If there are no errors with the words “Journal”, “watchDir” or “reindexing”:

"GetVolumePathName(dir, volume_path, MAX_PATH) failed in ChangeJournalWatcher::watchDir for dir "+dir
"CreateFile of volume '"+vol+"' failed. - watchDir"
"Change Journals not supported for Volume '"+vol+"'"
"Change Journals for Volume '"+vol+"' is being deleted"
"Error creating change journal for Volume '"+vol+"'"
"Unknown error for Volume '"+vol+"' after creation - watchDir"
"Unknown error for Volume '"+vol+"' - watchDir ec: "+convert((int)err)
"Journal id for '"+vol+"' wrong - reindexing"
"Last record not readable at '"+vol+"' - reindexing."
etc...

Also check that (after a client reboot) there are many more db lookups than filesystem lookups:

Indexing of “E” done. 1 filesystem lookups 8615 db lookups and 0 db updates

The client version should end with “-cbt” (e.g. 2.5.25-cbt). You can check it on the client’s about dialog, and on the server’s status page, by adding the “Client version” column.

Also, on the client’s about dialog, click “Show CBT status”, and you would see something similar to:

Volumes with active change block tracking and no crash persistence (v2.19):
C:, D:, E:, F:

If some volumes do not support CBT, they would be listed there as well.

Screenshot (359)

NOT CBT. My apologizes, I replied to another post about this just moments ago. Is CBT a paid upgrade or am I not looking in the right repo to grab the CBT version?

Indeed (link).

Thank you sir! :+1: