Vault Usage Report
This may be answered elsewhere but I could not find it, so forgive me.
Is there any way I can interrogate a Vault's internal database and obtain individual file sizes?
Reason I would like to do this, is that I want to know the 'worst offender' in regard to a full Vault.
I have created a new Vault on a different disk and started moving machines to it, but no sooner have I moved a machine, something fills it up again.
My life would be so much easier if I could isolate the biggest usage, and target the source machine as I should imagine it's probably useless file that don't even need backing up...
Thanks
Steve

- Log in to post comments

Hi Maria,
The Vault is located on a D: drive, seperate disks, dedupe is enabled and DB is with the Vault.
Vault total size is ~10TB, original size >120TB.
You may have misunderstood me. I know I can click on each machine and see the size, but I want to queries all files on all machines and find the Top 10 biggest files across all machines and backups within the Vault.
With this, I can either go to the machines and delete the large files or investigate exclusions in the backup job.
Thanks
Steve
- Log in to post comments

Hello Steve,
thanks for providing additional details.
Unfortunately, when the deduplication is enabled we can't even obtain the archive's size information except the manual browsing as I described above.
I want to queries all files on all machines and find the Top 10 biggest files across all machines and backups within the Vault.
This is an interesting suggestion. Thank you for this idea. I forward it as a feature request to the respective team.
- Log in to post comments

Maria,
Thats a real shame.
Is there not a way I can at least interface the ElasticSearch Engine with common SQL commands and query this data manually?
Or maybe the SQLite db3 files as I am used to querying those for the machines' last backup times.
I would just need to know what fils and view I should be looking at...
Steve
- Log in to post comments

Hi Steve!
I spent some time playing around with ElasticSearch since I'm also interested in accessing more low level data for querying.
Apparently it's quite easy to interface with ElasticSearch since it has a RestAPI. Though it seems the module that supports SQL queries isn't active/installed. However to get meaningful information out of it you'll have to dig deeper into how elastic search itself works and how it's configured. Reference here: https://www.elastic.co/guide/en/elasticsearch/reference/current/index.html
You can find the config for the ES endpoint in C:\Program Files\Acronis\CatalogBrowser\catalog_browser.json. For me it's localhost:9200 on the machine where the catalog service is installed. You can post/get commands there directly to the elastic search service with something like Postman.
I've managed to extract some info out of elastic search, though I'm not sure how accurate it is, but I've attached a query for searching files in all archives bigger than 1GB.
Some useful things I learnt with ES:
- Be careful with posts. You don't want to mess up the data
- /_cat will list categories you can query
- features/commands begin with underscore. With no underscore the text is interpreted as an index.
- It seems there is an ES index for each archive. (/_cat/indices?v)
Good luck.
Attachment | Size |
---|---|
509787-171377.png | 13.04 KB |
- Log in to post comments

Thanks for the information. Really helpful.
I have spent the day playing with ES and acrocmd list archive/backups to try and isolate my large files.
Given your 'size gte' example I can happily find the huge files which is great. What I am failing at at the moment is to then convert the '_index' guid to a friendly machine name.
I figured out the _index is actually 3 guid's seperated by an underscore (GUID_GUID_GUID). The second GUID actually relates to the Vault GUID. In my case I have 3 Vaults and looking at the records per GUID I could match GUID to Vault name.
Where I am stuck on is the 1st and 3rd GUID. Given that some of my 1st GUID's duplicate in the Vault, yet the 3rd GUID's are different, I am assuming that the 1st GUID=Machine and 3rd GUID=Disk.
Though I still cannot match either GUID's to the GUID's I find in exports from acrocmd list archive/backup...
It's all so close...
Looked at other _cat/s and not found anything to further me.
Will let you know if I find more.
Thanks again
Steve
- Log in to post comments

Péter,
The 1st GUID is definately the machine. If I open c:\programdata\Acronis\AMS\dml_resources.db3, table 'resources'. If I search for [id] = 'FirstGUID' it returns a single row with the machine hostname.
I have look through other db3 files looking to match the 3rd GUID, but found nothing useful as yet.
Steve
- Log in to post comments

Steve Smith wrote:Thanks for the information. Really helpful.
I have spent the day playing with ES and acrocmd list archive/backups to try and isolate my large files.
Given your 'size gte' example I can happily find the huge files which is great. What I am failing at at the moment is to then convert the '_index' guid to a friendly machine name.
I figured out the _index is actually 3 guid's seperated by an underscore (GUID_GUID_GUID). The second GUID actually relates to the Vault GUID. In my case I have 3 Vaults and looking at the records per GUID I could match GUID to Vault name.
Where I am stuck on is the 1st and 3rd GUID. Given that some of my 1st GUID's duplicate in the Vault, yet the 3rd GUID's are different, I am assuming that the 1st GUID=Machine and 3rd GUID=Disk.
Though I still cannot match either GUID's to the GUID's I find in exports from acrocmd list archive/backup...
It's all so close...
Looked at other _cat/s and not found anything to further me.
Will let you know if I find more.
Thanks again
Steve
Apparently the index is built like this: [resourceID]_[vaultID]_[archiveID].
Given that you have duplicates for the first GUID, you could double check this is indeed correct, since you most likely have a machine backed up to more than one vault, or you have multiple backup plans applied to a machine.
You can either find the archive in your backup location via that archiveID and backtrack to your machine, or search for the resourceID in the "backupmanager.resources" SQL table and read out the actual computername.
- Log in to post comments