'Server Error' on Windows client 7.0.4 and 7.0.9

Hi all - I’ve had an issue arise today on Windows desktop clients (both 7.0.4 and 7.0.9) accessing my Seafile 6.0.7 server running on a Windows 10 box.

The desktop client reports a “Server Error” and appears to have stopped syncing. I have restarted the server multiple times, as well as rebooted the PC the server is on and restarted, to no avail. I have also closed and restarted the client PCs, but am getting the same issue. I upgraded from 7.0.4 to 7.0.9 to see if that would fix it (doubtful, but wanted to try)…no luck.

The server itself seems to be working fine, as I can access it via the web and navigate as normal. I can also access using the mobile client on my iPhone, so I have to believe the issue is something in the clients not playing nice with the server.

Any thoughts? I did go into the client file logs and found the following repeating error:

[08/27/20 12:55:39] sync-mgr.c(582): Repo ‘My Library’ sync state transition from ‘initializing’ to ‘downloading’.
[08/27/20 12:55:39] http-tx-mgr.c(1157): Transfer repo ‘bad086e7’: (‘normal’, ‘init’) → (‘normal’, ‘check’)
[08/27/20 12:55:39] http-tx-mgr.c(1157): Transfer repo ‘bad086e7’: (‘normal’, ‘check’) → (‘normal’, ‘commit’)
[08/27/20 12:55:39] http-tx-mgr.c(1157): Transfer repo ‘bad086e7’: (‘normal’, ‘commit’) → (‘normal’, ‘fs’)
[08/27/20 12:55:39] http-tx-mgr.c(4010): Bad response code for GET server.name.com:1234/repo/bad086e7-6607-4516-bc0e-571f5bad67af/fs-id-list/?server-head=af6dee868753fd823f355220bc6395ecc8c51543&client-head=9c02877e60b60907fd247f9c8161bcb250e012ed: 500.
[08/27/20 12:55:39] http-tx-mgr.c(4580): Failed to get fs id list for repo bad086e7 on server server.name.com:1234.
[08/27/20 12:55:39] http-tx-mgr.c(1157): Transfer repo ‘bad086e7’: (‘normal’, ‘fs’) → (‘error’, ‘finished’)
[08/27/20 12:55:39] sync-mgr.c(621): Repo ‘My Library’ sync state transition from downloading to ‘error’: ‘Server error’.

Any help would be appreciated.

Thanks!

Jason

Did some Googling and found similar issues when there was block or file corruption on the server, typically during a hard power loss. I had network issues the other day before this happened, and am thinking something got garbled in a commit sequence. Ran FSCK on the server and came back with a number of corrupted files, which were repaired/replaced with 0 byte files. Re-sync’d my client files back to the server successfully, and all seems well again. This has, however, made me realize the importance of regular client backups, so a good lesson to ensure I have a good daily copy in case the server pooches again. :slight_smile: