@@ -1049,7 +1049,8 @@ components:
10491049 backupPact :
10501050 name : backupPact
10511051 in : query
1052- description : Flag, if present back up the Pact databases too
1052+ description : |
1053+ Flag, if present back up the Pact databases too. Extra disk space and time required
10531054 required : false
10541055 allowEmptyValue : true
10551056
@@ -2729,6 +2730,36 @@ paths:
27292730 /make-backup :
27302731 post :
27312732 summary : Start a backup job
2733+ description : |
2734+ Backup jobs are identified by the Unix timestamp when they're begun.
2735+
2736+ If a backup job is already in progress, this endpoint will return its
2737+ identifier instead of starting a new one.
2738+
2739+ The RocksDB portion of the database is always backed up; if backupPact
2740+ is set, the Sqlite (Pact) portion is backed up as well, taking much
2741+ longer.
2742+
2743+ There is no automatic backup retention policy - users need to delete old
2744+ backups.
2745+
2746+ This API is enabled by configuring the node thus:
2747+ ```yaml
2748+ backup:
2749+ api.enabled: true
2750+ directory: {some file path to put backups in}
2751+ ```
2752+ The backup directory ideally will be located in the same partition as the
2753+ RocksDB portion of the node database.
2754+
2755+ RocksDB backups to the same partition as holds the active RocksDB
2756+ database will have almost zero space overhead immediately, but over time
2757+ as the active database diverges from the backup the space overhead will
2758+ increase. If the backup is to another partition, it will take longer
2759+ and take as much disk space as the active RocksDB database.
2760+
2761+ Pact database backups always require about as much space as the active
2762+ Pact database does.
27322763 tags : [ misc ]
27332764 parameters :
27342765 - $ref : ' #/components/parameters/backupPact'
0 commit comments