Hi,
We recently suffered an outage at the data centre hosting Tableau Server on Ubuntu. It was brought back up but the backups started failing due to lack of space. We added an extra 25GB but it didn't help.
I ran some checks on the disk storage and found our /extracts folder is using 46GB:
77G /var
46G /var/opt/tableau/tableau_server/data/tabsvc/dataengine
43G /var/opt/tableau/tableau_server/data/tabsvc/dataengine/extract
However, according to the Server Space reports in Tableau, the entire server itself is only using 32B, of which about 20GB is extracts. Claude has suggested we have a bloating extracts folder because the extracts being refreshed during the outage were interrupted mid-write, and so haven't been removed cleanly on the next extract refresh.
I'm looking for a way of manually clearing out this folder to remove unnecessary space without affecting existing current extracts.
Thanks
#Tableau Server
orphan or duplicate extracts can occur but currently it is a just a theory in your case.
What Tableau Server do you have ? If >= 2025.1, you have recycle bin feature so you should try to empty it first.
- About orphans / duplicates, basically any hyper extract file present in dataengine > extract folder but not present in postgres repository table "
extracts " (identified by its "descriptor" value/folder, can be deleted.
- Similary any extract record reference present in postgres repository table "
extracts " with datasource_id null AND workbook_id null
, can be deleted from the dataengine > extract folder. But backgrounder is supposed to delete those automatically.
It is a it technical topic so take your time to audit your instance before doing any change and have a backup strategy in place.
And if any doubt, you can always open a Tableau Support case.
