Failed to assemble file on network error

[BobCaTT]

Hi

When a network error occur like the following, Newsbin may fail to assemble the file.
This is partially Ok because of par2 file, but only if they are present. Also, because of that failure, newsbin should gracefully retry.
I could also force the assemble incomplete and only one block was marked as missing.

Version 6.33.1882

Code: Select all

[07:36:08] ERROR  Socket  - Error: WSARecv Socket Error Code: 10060 A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.
[07:36:08] ERROR  NNTPSocket  - Server:xxx.xxx.xxx.xxx NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[07:36:08] ERROR  NNTPSocket  - SendCommand Worker - Server:xxx.xxx.xxx.xxxSendCommand - failed to Read Response From server: Socket - Read or Write Timeout


[DThor]

Under Options/Setup, what is Max Retries set to? I'm running the beta now, I don't recall offhand if that release had issues with files getting 'stuck' with disconnects, but Newsbin has had retries in it for a very long time.

DT

[BobCaTT]

Hi DThor,

The max retry is set to 2. Few things.
* The error is a bit random, but I had it few times already
* Only that connection had an issue while (7) others were running fine

[DThor]

At the risk of sounding like I'm putting you off, you might want to wait for next release before worrying about this for now. In the end there will always be network issues at one time or another, either the server end, your ISP, possibly your router, etc., and the important thing is that you have options to set that will determine how newsbin responds. Me personally, I have so few times now that this happens I tend to want newsbin to keep retries low and use aggressive mode, where if it gets stuck and there are enough pars online to just snag what it needs, repair and extract, the download waste be damned. If you're the sort of person that wants to preserve bandwidth but are never in a rush, you might want to set retries higher, it does what it sounds like. At the same time, you might want to hold off as I mentioned because I do recall things getting 'stuck' more frequently with that release than it does now.

The main takeaway is that I say things like 'stuck' like it's a bad thing, really it's doing what it should - you wanted something, it didn't complete for one reason or another, either you want newsbin to keep hammering at it till the end of time, or to wait for further intervention on your part.

DT