[tahoe-lafs-trac-stream] [tahoe-lafs] #1862: Triggering garbage collection

[tahoe-lafs]

#1862: Triggering garbage collection
-------------------------+-------------------------------------------
     Reporter:  tryb     |      Owner:  davidsarah
         Type:  defect   |     Status:  new
     Priority:  normal   |  Milestone:  undecided
    Component:  unknown  |    Version:  1.9.2
   Resolution:           |   Keywords:  garbage collection expiration
Launchpad Bug:           |
-------------------------+-------------------------------------------

Comment (by davidsarah):

 In the {{{[storage]}}} section of each storage server's {{{tahoe.cfg}}},
 set
 {{{
 expire.enabled = true
 expire.mode = age
 expire.override_lease_duration = 3600s
 }}}

 This setting is documented in [source:git/docs/garbage-collection.rst#L77
 docs/garbage-collection.rst], but the current documentation implies that
 no units shorter than a day are supported. Actually seconds are the
 default unit (or can be given as 's').

 '''However''', there is another obstacle to removing shares quickly, which
 is that the lease crawler runs no more than every 12 hours
 ([source:git/src/allmydata/storage/expirer.py#L49]). You can modify that
 line to make it run more often.

 When the leasedb changes are integrated (currently planned for Tahoe-LAFS
 v1.11.0), we can probably check for expired leases much more often (or
 potentially calculate what the next lease might expire and wait until
 exactly that time to re-check it). See #1833 for that.

-- 
Ticket URL: <https://tahoe-lafs.org/trac/tahoe-lafs/ticket/1862#comment:3>
tahoe-lafs <https://tahoe-lafs.org>
secure decentralized storage