All files missing in the Cloud but directory entries are there.

[Pete Graner]

Nothing has changed in over a week. Normal daily lucid upgrades. I went to add my new notebook to U1 today aka the 3rd machine and I clicked on U1 from the upper right applet (the one with online status) and selected U1 from the menu. It then took me to the U1 web interface, I logged in went thru the process of adding the computer. When that was one it took me to the U1 Web file listing and thats when I noticed all the files in the cloud were gone, however all of the directory structure is still there.

Binary package hint: ubuntuone-client

ProblemType: Bug
.home.pgraner..config.ubuntuone.syncdaemon.conf:
 [bandwidth_throttling]
 on = False
 read_limit = 0
 write_limit = 0
.home.pgraner..config.ubuntuone.ubuntuone.client.conf:
 [ubuntuone]
 connected = True
 connect = 2
 show_applet = 0
 bookmarked = True
Architecture: i386
Date: Fri Feb 26 16:02:40 2010
DistroRelease: Ubuntu 10.04
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Alpha i386 (20100224)
NonfreeKernelModules: nvidia
Package: ubuntuone-client 1.1.2-0ubuntu1
PackageArchitecture: all
ProcEnviron:
 PATH=(custom, user)
 LANG=en_US.utf8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.32-14.20-generic-pae
SourcePackage: ubuntuone-client
Uname: Linux 2.6.32-14-generic-pae i686

[Elliot Murphy (statik)]

During discussion on IRC pete has emailed syncdaemon logs to facundo for analysis, so assigning to him.

[Facundo Batista (facundo)]

Analysis of the logs:

First machine (http://paste.ubuntu.com/384641/), where all was ok, but then
it started to delete files.

Log 1 - 2010-02-17 14:44: SyncDaemon started. Did Local Rescan, finished it,
        and logs stop abruptly as if the process was killed.

Log 2 - 2010-02-17 14:45 to 2010-02-18 14:43: SyncDaemon started. Did
        Local Rescan, then Server Rescan. Nothing strange, but it got
        stuck in START_WORKING_ON_BOTH (known bug, fixed last week),
        and then because a network hiccup it got stuck in
        START_STANDOFF_WAITING_WITH_NETWORK_WITH_BOTHQ (also a known bug,
        #487257, work in progress right now).

Log 3 - 2010-02-18 14:45 to 2010-02-19 14:43: still stuck in
        START_STANDOFF_WAITING_WITH_NETWORK_WITH_BOTHQ.

Log 4 - 2010-02-19 14:45 to 2010-02-20 14:43: still stuck in
        START_STANDOFF_WAITING_WITH_NETWORK_WITH_BOTHQ.

Log 5 - 2010-02-20 14:45 to 2010-02-21 09:45: still stuck in
        START_STANDOFF_WAITING_WITH_NETWORK_WITH_BOTHQ.

Log 6 - 2010-02-26 13:58 to 2010-02-26 15:22: SyncDaemon started (five days
        later).

        Local Rescan started, and finished (lot of work here, ten thousand
        log lines in eight seconds, lot of new or modified files/dirs,
        but NO DELETION generated here).

        Server rescan starts 35 minutes later (meanwhile, a lot of files were
        shared). It queries 1069 objects. Then, iot starts to receive a lot
        of deletes from the server, zillions of SV_FILE_DELETEs.

Second machine (http://pastebin.ubuntu.com/384568/):

Log 1 - 2010-02-26 13:25 to 13:49: SyncDaemon started. Local rescan started
        and finished. Never received order to connect from user, so even
        having network, it didn't touch the server.

Conclusion:

Let's first discard stuff, to then focus easily on the core of the problem.
In the first machine until the first log, nothing really happened, and in
the second machine we see that it never got into the server, so it could not
delete anything.

The log number 6 of the first machine has a some things to consider:

- No deletion was generated from the client: No delete file or dir events
  were generated (by syncdaemon or user).

- After the server rescan, it received 1582 SV_FILE_DELETE: this tend to
  confirm that the deletion came from the server.

- The client, at ~14:35:10, received from the server that the free space was
  2147483648, which is exactly 2GB. This is another indication that no files
  exist in the server.

So, I can not see in these logs any indication of why those files
disappeared, beyond the reassure that the deletions were not generated in the
client.

Maybe authentication errors have something to do? User reports that after
trying to setup the second machine it saw that the web ui stopped showing
files, and then is when the authentication errors started.

So, Elliot, I'm assigning back this to you, to see if we can continue this
analysis at other level.

*Please*, let me know if I can help you somehow.

[John O'Brien (jdobrien)]

*Please*, let me know if I can help you somehow.
> ** Changed in: ubuntuone-client
> Assignee: Facundo Batista (facundo) => Elliot Murphy (statik)
>
>
Just an fyi, the quota stats for this user show 251.2 kb used and 450MB
of deleted data.

[Facundo Batista (facundo)]

John, *exactly*.

So why suddenly server starts to say that it has the full 2GB, and that files are not there any more, and auth fails? Maybe we're dealing with two accounts? Can we check this somehow?

[Pete Graner (pgraner)]

For clarity I have three machines:

1. zorak
2. desktop
3. air

zorak & desktop were both up and running with U1 and syncing happily. I was adding air a new notebook when I noticed the problem. I have provided the logs from desktop & air below are the pastebin URLs for zorak:

http://paste.ubuntu.com/386294/
http://paste.ubuntu.com/386295/
http://paste.ubuntu.com/386296/
http://paste.ubuntu.com/386297/
http://paste.ubuntu.com/386298/
http://paste.ubuntu.com/386299/
http://paste.ubuntu.com/386300/
http://paste.ubuntu.com/386301/

[Eric Casteleijn (thisfred)]

Facundo, Elliot, do these logs tell us enough to mark this confirmed/triaged?