Cant speak for other Plattforms but on Windows the Icon State of the Seafile Client goes to error
if a simple disconnect happens.
Like a daily Reboot for Backup leaves it always in error in the Morning.
This then requires manual clear by at least looking at that error.
Id rather see the client honor the current State. So if the error (like isconnect) is already resolved the icon should reflect that.
To me its just annoying, but to a regular user it seems like the client is never really working.
The client should adapt the Icon to the real state of the application without userinteraction
Microsoft for example does this (onedrive) and even supresses Shortterm errors (outlook and onedrive) if the error is resolved soon after.
Hello @Terrorfrog ,
Are you using the latest version of the seafile client? Older versions may experience persistent network errors, which can cause the icon status to display incorrectly.
yes i do,
and i already stated that once errors are cleared it stays clear until the next service reboot / short network interruption
Hello @Terrorfrog ,
After restarting the client, what error message is actually displayed when you check the sync errors? If possible, could you provide a screenshot?
No as described after restarting the Server
So client runs, server makes a daily reboot to backup in an offline state.
When its back online seafile shows error until manually cleared.
Same things happen when connection is interrupted for a short time.
I have the same issue on my Mac. Every time the server connection is down, the exclamation mark pops up, and it is ever only resolved by going to the “Show file sync errors” screen. And it has been this annoying way for years.
(One of the two reasons I would never recommend Seafile to non-pro users. The other one is the missing auto-update feature which is just bonkers IMHO.)
i want to argue it is the hole point of using a sync solution instead of a plain smb share to expect
changing connection state. like on laptops during travel etc…
so i think even the error is wrong on no connect to server. it should say offline until online again.
or paused.
only on real errors - aka cant sync but server is online and reachable - should an error be presented.
while it loos minor, it makes a hole different perception for users (and evaluation testers)
the current reporting simply make it look very fragile
i also agree with the autoupdate. i see that seafile wants to target corporate customers and there autoupdate WAS less of importance as most manage their own update cycles via tools.
However this changed a lot in recent years, with auto provisioning and managed in the cloud
this kind of im premise tools are a lot less relevant and no built in update function (ideally via policy like set version via on group level or similar) is a PIA
for SMB customers no autoupdate is annoying, if they are a bit bigger (20+ seats) its probably a breaking missing feature for many
Hello,
We will improve some error messages caused by network issues in the next release.
Uhh that wording is scary.
It is not so much that error messages need improving (thats too but thats another cup of coffee)
its about that there should be no error to begin with.
Server not reachable (its a state of the connection, not a verified error)
-Client Offline / Paused - which would be the correct reflection of the actual state
-should heal itself when connection is reestablished aka toggle between Online and Offline depending on the actual state
Server returns an error (whatever that maybe)
Now we have a valid and VERIFIED error state
-client should display error if its persistent
-may or may not heal itself thats the question what errors should heal themself and which need manual review. in theory all temporary errors should heal themself but thats something only you guys know for certain
Keep in mind the client is on the user side. he aint gonna debug (or should even try).
all the user needs is a proper easy to decifer information of
-is online everything honky dory
-is offline oups maybe reconnect more wireless cable
-error - call tech support
he doesnt need to know what was yesterday, he needs to know what is NOW.
is everything fine now or it isnt…
1 Like
Hi,
The issue you’re experiencing mainly occurs because the seafile server has been stopped while nginx is still running. In this situation, when the client sends requests to nginx, it returns a 502 error. From the client’s perspective, a 502 error is treated as a server-side error, so the client records it and notifies the user.
Under normal circumstances, the server is not restarted frequently, so the client would not receive such notifications often. In the next version, we will optimize this behavior to provide users with a more reasonable and user-friendly explanation of these errors.
This is not the case at all ant logs show never a 502 error.
I usuall get libcurl failed to post (or get)
with a subsequent either server timeout or couldnt connect to server
same error pattern (and failure icon) i get when networkconnection gets interrupted.
the backup doesnt shutdown seafile itself and leave nginx running.
the entire VM gets shutdown including all docker stacks, snappshotted and fired back up.
and the error pattern isnt just after backups, its at any network interruption
..
Besides user dont need an explanation. is it so hard to understand ?
the user will not understand either way. the user needs an easy way to understand what he current state is. not why , or what it was 3 hours ago.
the icon should reflect
- running
- paused
- in permanent error needs intervention (aka support)
not an explanation why, this is explained in the logs for people who are supposed to get an explanation and know what todo about it