Does the Stellar network have the concept of pruning? If so, how do I go about configuring it?
I am running a node for development purposes but don't want the database and disk usage to become larger than a couple of gigs. Is that possible?
Add a comment
|
1 Answer
It looks like you are looking for the /maintenance HTTP command. It cleans unnecessary Stellar Core history, including XDR buckets stored on disk and records in the database.
If you plan to use Horizon in conjunction with Core node, you will only have access to the recent Stellar history, as a complete history archive requires a fully-synced Horizon node (currently more than 50 GB for public network).
answered Aug 10, 2018 at 0:18
-
I'm interested in this as well. Could you elaborate a little more about it? Is it taking old ledgers and simply deleting them? Or is it transferring those ledger data to an archive? When I run the
/maintenancecommand on a node connecting to the test network, I get[default INFO] No work performed.– B.LiOct 6, 2018 at 0:30 -
Probably you have
CATCHUP_COMPLETE=falseconfig parameter set. The/maintenancecommands removes stale ledgers retaining only lastCATCHUP_RECENTledgers and performs/bucketsdir cleanup. You can configure automatic maintenance task runner usingAUTOMATIC_MAINTENANCE_PERIODandAUTOMATIC_MAINTENANCE_COUNTconfig params. Oct 6, 2018 at 1:13 -
I just tried it both on nodes with
CATCHUP_COMPLETE=falseandCATCHUP_COMPLETE=truewithout touchingCATCHUP_RECENT. Both still produces the same result.– B.LiOct 6, 2018 at 1:25 -
1Set
CATCHUP_COMPLETE=falseand just wait for some time. The/maintenancecommand only schedules a maintenance job. When I run/maintenanceon my node there were no changes for quite some time. I even restarted the service a few times. And then, two hours later, it started to delete records from the database in huge batches. Don't know the internal mechanics of the process, so can't tell exactly how the maintenance is scheduled. Oct 6, 2018 at 2:05