Up to date but not syncing

Support
1

I’m using the current master f05ccd775abf1a51a39b0363d72fe7942815e01b and I have no idea why it’s not working. I already reset both databases and it seemed to work a while ago. They also should have quite similar files.

Also the out of date dialog doesn’t seem to work at raspberry pi’s side:

image
image2667×1217 251 KB

raspberry pi (Arch Linux ARM):

image
image2667×1217 260 KB

laptop:

image
image2667×1217 265 KB

Logs: https://gist.github.com/mohe2015/acb47cbc105d198d6a4e055d27a27f95

I either think it’s a network issue or some kind of deadlock.

Any help would really be appreciated. If you need anything else just tell me.

2

This looks more like a problem with the new need metadata accounting: The numbers in the remote devices section are based on that and the actual list is obtained directly from db. In your case they apparently don’t match. That change was introduced in v1.6.0-rc.1 and I don’t have such problems. And unless you find a way to reproduce it I don’t see how to debug this. The relevant debug facility for this is db.

If you run with STRECHECKDBEVERY=1s, does the problem go away?

3

Yes, running with STRECHECKDBEVERY=1s seems to have fixed the problem. I just restarted without that and found some suspicious log message:

[BRQHQ] 14:20:16 INFO: Stored folder metadata for "cubot_note_s_q4ha-photos" is 441518h20m16.019778363s old; recalculating

Full logs: https://gist.github.com/mohe2015/e27a8d0e1805e122a449e2fd366fb1f1 (unfortunately without debugging info)

Also I have multiple megabytes of logs while running with STRECHECKDBEVERY=1s and db logging so I could provide them if they could help find the problem.

4

Sorry about that, the two things were meant separately: If you can get the problem to appear while having STTRACE=db active, that would be helpful. Running with STRECHECKDBEVERY=1s is to try and get rid of the problem (which I am glad to hear it did), no need for logs there, debug or not.

I assume that’s the time to 1970: We set the created time to 0 if we want to force recreate metadata. Which should probably be handled more nicely on a user visible log line but nothing to worry about :slight_smile:

5

Thanks a lot for the help and all the explanations. I may run with STTRACE=db for a while as long as it doesn’t impact performance that much and “hope” the problem shows up again.

Feel free to close this or maybe leave it open if somebody else also gets this problem.

6

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.