Nyuu just hangs at 99%, sometimes even at 100% #99
Comments
|
Thanks for the report + info. Could I get you to confirm the version of Nyuu you're using, as well as the error that occurred (look for |
|
I'm using 0.4.1 Could you please tell me the location of the log? I don't really know where it is Also, it's happened again. This is nyuu output from just now |
|
Well the log will likely vary by run, so it's possible you don't get an error on some runs. Would you happened to have changed the Perhaps I can see if there's a way to make it more obvious that it's stalling on checks. Otherwise, you can use the |
|
This is a part of my config. I will be using the --verbose option upon next try Is there a way to tell it to just finish if it tested and found, say, 99% of the post already? 99% w/ pars should be pretty much safe to assume the posts is up 100%. It seems to hang on not finding 1 (out of sometimes thousands of) article. I use config.js for most of the config options Servers : These are my check options |
|
With multiple servers, actions are currently done randomly against them, so there's no guarantee that checks are performed on the same server that posts are. Yeah I thought you may have changed those options - basically that says that if a post isn't found via check, wait 100s and check again. Repeat this up to 4 times, and repost if not found. After reposting, repeat this whole check process - meaning that an article not accepted by the host could hold up the process for quite a long time.
That's a neat idea - I'll look into that. |
|
I have kept a close eye on the primary issue that happened, and it hasn't occurred again yet. I can't recall having made any changes to the main config. When it does, I'll report again, hopefully with better evidence Edit: I actually did add a check connection to each server. I have noticed some servers do eat up bandwidth even when only posting, which is not optimal for blocks long run, which is why I disabled it in the first place |
|
It happened again. This is the output of the web server: I haven't ended the process yet. It's reporting inactivity for a long long time, much longer than in the settings. in terminal I see |
|
Not sure what the cause is there :/ Are the settings largely the same as those above? Otherwise, I'll look into this a bit more, and beef up diagnostics/info to help track it down. Thanks for the update. |
|
I've added more info to the status server if you visit If you're still getting random stalls, the info shown there might help. |
|
Thank you for the update. Upgraded local install to the latest commit yields the following results: And then it shows as So far so good, I'll keep testing it out to see where/if it hangs |
|
It hasn't happened in awhile, but today just had the first hang-up after the update This time there is no "extra info" shown, only no ETA I have to say that this is the very first time it happened since updating, and I have been testing it quite exhaustively |
|
Thanks for all the testing + info! It looks like a post got stuck somewhere, but I can't think of a cause. Are you able to post a full log? The details suggest that errors occurred, but your posted log doesn't show what they are. |
|
It's possible I may have jumped the gun on that one, I think it got stuck and was retrying. It only happened that one occasion though. As we have established, I have liberal retries because my connection sucks. I tried again with a much longer queue and left it to do it's thing, and haven't found any messages of failure yet. I think that one example was because I was scrutinizing it too closely and caught it before it could handle the erroring gracefully mostly because of my liberal error-handling settings. I think everything works nicely as it's intended right now Edit: I will do the |
|
Thanks for the info. |
|
Changed the status screen to show a bit more info - hopefully that helps with tracking things down better. Let me know if you encounter the issue again with the new code. Thanks! |
|
This only happens on a slow connection. I haven't seen a single fail since upgrading speed. I don't understand why, but I've literally had 0 incidences since. I think you can close this ticket, even though the actual reason why it failed we don't know yet. |
|
I'm running into hanging issues. using 0.4.1 HTTP debug server output: config is default. when opening the nzb you can see that one of the parts is missing: 731/732 edit: another hang on the upload directly after: HTTP server: |
|
Thanks for the detailed info @duhagppn ! I can see the cause of your first example. A skipped I'll look at implementing a fix for the stall. Regardless, you can avoid connection rejections by lowering your connection count.
|
Requires post checking to be active, 'connect-fail' error skipping and some working connections Ref #99
|
I'm fairly certain this is an issue with vipernews as well, could you try with a different provider @duhagppn ? |
|
was confident this was fixed but woke up to anotherr hung process. these are the only posting errors: |
|
Thanks for the report - that's quite weird, so I'll investigate more. From the looks of things, at least that post seems like it did complete. |
|
I believe it's a ViperNews issue. I removed ViperNews from my servers and I haven't had this problem again. Something is going on with ViperNews @duhagppn Could you please try with a different server and report back? |
although it reached 100% the post was incomplete - there was 1 file completely missing from the nzb. 20 .par2 index files, 19 media files, 140 .par2 volumes + 1 nfo. I did kill the process once it was clear the upload was stalled, but surely what was already posted should've been written to the nzb? @df-a I don't have any other posting accounts to try with, certainly not at 15 eur/20 usd a block like farm or usenet express 🤣 personally I think the issue is less with vipernews specifically and more to do with how nyuu is handling posting/checking/retries when encountering things like authentication or timeout issues. in the previous post log for example, you can see Connection #15 is the only one with "finished" where others are listing their transfer & idle stats etc. is there a combination of the log & progress options to output standard info to the terminal while writing debug output to a log file? |
…the check queue Previously, post-checks would get lost if the connection failed Ref #99
Ah, that's unfortunate. I think I've found the issue - if a check connection gets lost, the post it's checking also does too, so the process doesn't complete. This should be fixed in the latest code. The NZB is written as the upload is progressing, but keep in mind that it has to be done in order, and can't "go backwards". If post checking is enabled, segments can only be written once the check is successful, since a failed check may result in a re-post, which could mean a different message ID.
No, though the debug info isn't particularly useful. The progress server is currently the most useful for these types of issues. |
|
thanks, I've updated to the latest version and will let you know if it occurs again. fingers crossed it's all fixed |
Hello,
Nyuu sometimes randomly hangs at 99%, sometimes even at 100%. It doesn't happen all the time.
I have
skip-errors: true. I also have multiple servers configured, yet only one is configured for checking the articles posted.This is what the built in web server reported after one recent such hang:
When this happens, I keep seeing
NNTP connection lostechoed multiple times, but it doesn't finish the postThe text was updated successfully, but these errors were encountered: