Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Client trying to sync deleted File #2919

Closed
Turbocube644 opened this issue Mar 4, 2015 · 26 comments
Closed

Client trying to sync deleted File #2919

Turbocube644 opened this issue Mar 4, 2015 · 26 comments
Assignees
Labels
bug p2-high Escalation, on top of current planning, release blocker ReadyToTest QA, please validate the fix/enhancement

Comments

@Turbocube644
Copy link

The Client (1.8.0 beta1) tries to sync a file, which has been deleted. It shows the following error: "Error downloading [filename] - Server replied: Not found". The file neither exists on the server nor the local computer. This happens for multiple files, that have been deleted.
For some classes, I often get a Timeout.
The server ist still 7.0.4 (because the update routine thinks it is the newest version)

@guruz
Copy link
Contributor

guruz commented Mar 5, 2015

@Turbocube644 Is there anything about the filename or the directory it is stored in that could help us to track this down?

Special characters?
External storage on server?
Which DB on the server?

@guruz guruz added bug p2-high Escalation, on top of current planning, release blocker labels Mar 5, 2015
@guruz guruz added this to the 1.8 - UI Enhancements milestone Mar 5, 2015
@Turbocube644
Copy link
Author

The file name only consists out of ASCII letters, so no special characters.
The directory is no external storage, but the file is part of a git repository.
I'm using a mysql database.

@guruz
Copy link
Contributor

guruz commented Mar 5, 2015

@icewind1991 @PVince81 A server file scanner issue?

@PVince81
Copy link
Contributor

PVince81 commented Mar 5, 2015

Can you check the database: select * from oc_filecache where path like '%thedeletedfilename%' ?

@Turbocube644
Copy link
Author

I checked the database: the query returns zero entries.

@PVince81
Copy link
Contributor

PVince81 commented Mar 5, 2015

@guruz seems not. If the file is neither on disk nor in the oc_filecache, the server should not know anything about it and return 404.

How would the client behave ? What if it's trying to resume a failed download on a file that was deleted ? (just wild guess)

@guruz
Copy link
Contributor

guruz commented Mar 5, 2015

What if it's trying to resume a failed download on a file that was deleted ?

Should not happen here, if the file was not discovered the download will not be attempted to resume.
CC @ogoffart

@Turbocube644 Are you 100% about that DB query results? Also about if the file maybe exists on your server hd?

@Turbocube644
Copy link
Author

Yes, I'm 100% sure about the DB query results. I checked them twice.
I just had a look at the server hd: the file does not exit there.

@dragotin
Copy link
Contributor

dragotin commented Mar 9, 2015

Do you find an entry about the file in the apache access_log? That entry would be interesting.

@Turbocube644
Copy link
Author

Unfortunately I don't have access to the logs...

@dragotin
Copy link
Contributor

@Turbocube644 ok, please create a client log file that shows a sync run where your client tries to download the files.

@Turbocube644
Copy link
Author

@ckamm ckamm assigned ckamm and unassigned dragotin Mar 11, 2015
@ckamm
Copy link
Contributor

ckamm commented Mar 11, 2015

@Turbocube644 Thanks for the log! That's interesting.

The whole sync failed because of a timeout when trying to delete workspace/.metadata/.plugins/org.eclipse.core.resources/.history/6b/b0c2edc09db40014111b916fb9f5489a from the server. That file is found on the server - is that correct?

The confusing messages saying 'Error downloading ...' are actually about DELETE requests to the server failing.

But then there's also a file ending in c06771af7ab50014115594bbc2625284 that shows up only in the remote reconciliation logs (why is it assumed to be on the server?). Marking this as REMOVE and trying to delete it is bound to cause trouble.

@ckamm
Copy link
Contributor

ckamm commented Mar 11, 2015

The etag for the directory that contains c06771af7ab50014115594bbc2625284 didn't change, which is why the remote discovery assumes the file is still present and attempts to delete it.

@ckamm
Copy link
Contributor

ckamm commented Mar 11, 2015

We're considering whether to make 404 on DELETE a non-fatal error that deletes the file from the DB. But it won't be in 1.8.0.

@ckamm
Copy link
Contributor

ckamm commented Mar 13, 2015

With the patch applied I no longer get stuck files in this situation.

@guruz
Copy link
Contributor

guruz commented Mar 13, 2015

The etag for the directory that contains c06771af7ab50014115594bbc2625284 didn't change, which is why the remote discovery assumes the file is still present and attempts to delete it.

Hold on.. something for @PVince81 ?

@guruz
Copy link
Contributor

guruz commented Mar 13, 2015

@dragotin OK to cherry-pick the above commit into 1.8 for rc2? It's a gold ticket

@guruz guruz assigned guruz and unassigned ckamm Mar 13, 2015
@PVince81
Copy link
Contributor

Is it a file inside a shared folder ? Is the user seeing the bug a recipient or owner ? Is the shared folder inside a subfolder ? If yes, it could be the cross-storage etag propagation issue.

@dragotin
Copy link
Contributor

Yes, please pick the patch now.

@dragotin dragotin assigned ckamm and unassigned guruz Mar 18, 2015
@ckamm
Copy link
Contributor

ckamm commented Mar 19, 2015

Merged into the 1.8 branch now.

@ckamm ckamm closed this as completed Mar 19, 2015
@ckamm ckamm reopened this Mar 19, 2015
@ckamm ckamm added the ReadyToTest QA, please validate the fix/enhancement label Mar 19, 2015
@ckamm
Copy link
Contributor

ckamm commented Mar 19, 2015

@Turbocube644 The patch was merged, could you test it with tomorrow's build at http://download.owncloud.com/desktop/daily/ ?

@guruz
Copy link
Contributor

guruz commented Mar 19, 2015

@Turbocube644 Could you still reply to the questions in #2919 (comment) ?
Thanks

@Turbocube644
Copy link
Author

No, it wasn't inside a Shared folder.
Unfortunately my owncloud instance completly crashed while updating to 8.0.2 and I therefore had to reset completly (the backup didn't work). Now After resyncing I didn't get the problem again.

@ckamm
Copy link
Contributor

ckamm commented Mar 25, 2015

Okay! Closing since the patch fixed it when I reproduced the problem.

@ckamm ckamm closed this as completed Mar 25, 2015
@rienesl
Copy link

rienesl commented May 3, 2015

Just to give an update:
I have the same problem. Having updated to 8.0.2 on the server side helped nothing. I'm using the 1.8.0 client. It happens, after I renamed a directory. Since then, I get the error on all the clients, . Pressing F5 for rebuilding the index did not help. I'm using a productivity system, so I'm not able to try the daily-built test.
My suggestion would be: If a deletion request returns "not found" then remove it from the database (the goal of deleting it was to remove it anyway...).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug p2-high Escalation, on top of current planning, release blocker ReadyToTest QA, please validate the fix/enhancement
Projects
None yet
Development

No branches or pull requests

6 participants