Interrupted downloads in a read only share never are resumed
Bug #598189 reported by
Facundo Batista
This bug affects 2 people
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Ubuntu One Client |
Fix Released
|
High
|
Guillermo Gonzalez | ||
ubuntuone-client (Ubuntu) |
Fix Released
|
High
|
Ubuntu One Foundations+ team |
Bug Description
When the download is started, the server_hash is updated in the client.
Suppose Syncdaemon is stopped, and restarted.
LR does not check RO shares, so nothing is found there, and SR sends the same hash that the server has, so the interrupted download never is resumed.
The solution is for LR to check all metadata for RO shares (not disk), and restart downloads if it finds nodes in SERVER state.
Related branches
lp:~verterok/ubuntuone-client/fix-598189
- John O'Brien (community): Approve
- Facundo Batista (community): Approve
-
Diff: 122 lines (+94/-0)2 files modifiedtests/syncdaemon/test_localrescan.py (+72/-0)
ubuntuone/syncdaemon/local_rescan.py (+22/-0)
Changed in ubuntuone-client (Ubuntu): | |
importance: | Undecided → High |
status: | New → Confirmed |
tags: | added: foundations+ |
Changed in ubuntuone-client: | |
assignee: | nobody → Ubuntu One Foundations+ team (ubuntuone-foundations+) |
Changed in ubuntuone-client (Ubuntu): | |
assignee: | nobody → Ubuntu One Foundations+ team (ubuntuone-foundations+) |
Changed in ubuntuone-client: | |
status: | Confirmed → In Progress |
assignee: | Ubuntu One Foundations+ team (ubuntuone-foundations+) → Guillermo Gonzalez (verterok) |
Changed in ubuntuone-client: | |
status: | In Progress → Fix Committed |
Changed in ubuntuone-client: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
ubuntuone-client (1.4.1-0ubuntu1) maverick; urgency=low
* New upstream release: subscribe_ udf method to handle the error cases properly subscribe_ udf method (Guillermo
- Fix VolumeManager.
(Guillermo Gonzalez) (LP: #632439)
- Make ubuntuone-launch work with the new sso (John Lenton) (LP: #631822)
- Wait for READY before asking for connection (John Lenton) (LP: #525743
and LP: #518027)
- Move operations live in limbo until confirmed (Facundo Batista)
- Fix a few problems with emblems (Rodrigo Moya) (LP: #607247)
- Load UDFs as soon as syncdaemon is ready (Rodrigo Moya) (LP: #612541)
- Load public files as soon as syncdaemon is ready (Rodrigo Moya)
(LP: #627339)
- When VM receives an event on a volume it doesn't have, refresh them
(Facundo Batista)
- Made u1sync client work from the command line. No more "authorize" option,
only --oauth (Natalia Bidart) (LP: #634337)
- Marker chains are errbacked ok now (Facundo Batista)
- Remove multiple returns in VolumeManager.
Gonzalez)
- Do calls to libproxy in a separate process (John Lenton) (LP: #633241)
- OAuth consumer key and secret can be passed as parameter (Natalia Bidart)
(LP: #637022)
- Don't generate a new file if we're deleting or overwriting it
(Facundo Batista)
- Make sure that the u1 music path is present (Manuel de la Peña)
(LP: #558762)
- Make the "Current plan" label show the actual value returned from the
server (John Lenton) (LP: #551138)
- Add --with-sso option to configure, for building with uninstalled
sso-client (Rodney Dawes)
- Fix volume manager to create the fsm node on server rescan if the node
isn't there (Guillermo Gonzalez) (LP: #632454)
- Fix Local rescan to process read only shares and resume interrupted
downloads (Guillermo Gonzalez) (LP: #598189)
- Make sure we load list of shared folders in Nautilus plugin (Rodrigo Moya)
(LP: #633013)
- Dereference the possible markers in limbos (Facundo Batista)
- Don't start syncdaemon from Nautilus plugin. Just wait for it to be
available (Rodrigo Moya) (LP: #613219, LP: #635629 and LP: #628877)
-- Rodrigo Moya <email address hidden> Wed, 15 Sep 2010 22:17:37 +0200