Getting Server Error 500 on some libraries (Pro Version)

Hello,

We are getting server error after upgrading to pro server 8.0.3 from 7.1.8. Some libraries are working some are getting server error and didn’t sync. I tried to delete the library, then make seaf-gc and seaf-fsck and recreate and empty library with same name and still getting the same error.

I tried to look at the server logs i couldn’t find anything usefull. I only find the error below on client log:

[05/07/21 17:35:28] http-tx-mgr.c(4010): Bad response code for GET https : /xyz.xyz.com/seafhttp/repo/df90e78f-3a85-4a6a-963c-9210a0caa3b2/fs-id-list/?server-head=f6dceddaa74cf6f24fa8bb276943af86e5246773: 500.
[05/07/21 17:35:28] http-tx-mgr.c(4590): Failed to get fs id list for repo df90e78f on server https : /xyz.xyz.com.
[05/07/21 17:35:28] http-tx-mgr.c(1157): Transfer repo ‘df90e78f’: (‘normal’, ‘fs’) → (‘error’, ‘finished’)
[05/07/21 17:35:28] clone-mgr.c(697): Transition clone state for df90e78f from [fetch] to [error]: Server error.

If i try to open the url from browser, it gives 500 error. But we have no problem before 8.0.3. On update there was no error occured.

Is there a way to find the cause of the problem and fix?
Or is there a way to rollback to version 7.1.8?

Edit 1: I don’t know why but we have no seahub.log file on the server.

Edit 2: I’ve also set ‘max_sync_file_count = 1000000000000’ on seafile.conf file.

Thank you

Now i downgraded to 8.0.2 and everything seems working. I think that there is a problem with “max_sync_file_count” parameter on the new version.

@daniel.pan

the same, I have exactly the same problem with version 8.0.3 pro, since the activation of parameters in seafile.conf:

[fileserver]
max_sync_file_count = 1000000000
fs_id_list_request_timeout = 86400

when going back to version 8.0.2 pro no syncronization problem
Is there a way to disable this setting please ?
debian 10, MariaDB

@ben
when I apply this following configuration with seafile pro 8.0.3:

max_sync_file_count = 1000000000
fs_id_list_request_timeout = -1

I no longer have an internal server problem with large libraries …

Hi, I’m having exactly the same problem, also throws the same error in the log. I tried changing the max_sync_file_count and fs_id_list_request_timeout in my /opt/seafile/conf/seafile.conf like @Cisco suggested, but when I ran seafile.sh restart, it seemed like Seafile simply ignored my configuration, because it logged this:

2021-05-13 19:46:28 http-server.c(428): fileserver: fs_id_list_request_timeout = 300
2021-05-13 19:46:28 http-server.c(441): fileserver: max_sync_file_count = 100000

I’m on Seafile Pro Server 8.0.3. on CentOS 8. I installed it with the installer script. Also, I hadn’t added max_sync_file_count and fs_id_list_request_timeout in the settings before, the error occured with the standard settings, fresh install.

I could work around the issue from the client side: Apparently this happens when you try to sync an empty (on the server) library with an existing non-empty folder on the computer. (At least this seemed to be the problem for me, I don’t know if you can reproduce that.) I could solve it by uploading a single file to the library manually via the web interface and then syncing the library via the client. Pretty annoying bug as it seems.