Comment 6 for bug 476711

In order to add more data to make a decision, I've been gathering logs from people to analyze the trade off between metadata loading time and local rescan running time.

For a "real" user on #ubuntuone (psyphercode), from http://paste.ubuntu.com/393987/

2010-03-12 07:15:59,194 - ubuntuone.SyncDaemon.fsm - INFO - loading updated metadata
2010-03-12 07:54:56,636 - ubuntuone.SyncDaemon.fsm - INFO - initialized: idx_path: 195799, idx_node_id: 14082, shares: 1
2010-03-12 07:54:56,642 - ubuntuone.SyncDaemon.HQ - INFO - HashQueue: _hasher started
2010-03-12 07:54:57,239 - ubuntuone.SyncDaemon.DBus - INFO - DBusInterface initialized.
...
2010-03-12 07:54:57,241 - ubuntuone.SyncDaemon.Main - NOTE - Local rescan starting...
2010-03-12 07:54:57,241 - ubuntuone.SyncDaemon.local_rescan - INFO - start scan all shares
2010-03-12 08:05:48,303 - ubuntuone.SyncDaemon.Main - NOTE - ---- MARK (state: READING; queues: metadata: 0; content: 0; hash: 0, fsm-cache: hit=8 miss=587391) ----
...
2010-03-12 08:10:24,726 - ubuntuone.SyncDaemon.Main - NOTE - Local rescan finished!
2010-03-12 08:10:24,727 - ubuntuone.SyncDaemon.Main - INFO - hash queue empty. We are ready!

 * meatdata loading ~ 40 minutes
 * local rescan ~16 minutes