Server error causes incomplete downloads

[qinglung2]

It seems like a server error sometimes causes all the threads to disconnect. This is not specific to alt.binz, as I've seen it happen in Grab-it as well. However, where Grab-it would just retry the posts that got interrupted, alt.binz goes ahead and writes the incomplete posts to disk and carries on to the next posts, sometimes resulting in incomplete files which even the PAR2 files can't correct.

[cr4zyfr4g]

you can set retry count in Setup -> Connection, that should solve your Problem.

[Rdl]

It seems like a server error sometimes causes all the threads to disconnect. This is not specific to alt.binz, as I've seen it happen in Grab-it as well. However, where Grab-it would just retry the posts that got interrupted, alt.binz goes ahead and writes the incomplete posts to disk and carries on to the next posts, sometimes resulting in incomplete files which even the PAR2 files can't correct.

That looks so much like Hitnews. But not retrying, that's not true. I would like to see detailed log proving that.

[qinglung2]

It does retry when there is a timeout... just this certain error when all threads are dropped that it is not retrying.

I have seen some error message when it happens, I'll try and get it to show up in the log again and post it. Other times it happens and there is no error message, very strange.

[qinglung2]

All the logs really say is: "StatusCode: 0", so one thread is disconnecting and then I can see all the others disconnect at the same time, though the log says "article finished" for all the other threads.

When I look in the cache folder, I can see that the posts that were written after the disconnects are not the same size as the completed ones.

Is there no way for the log to give more details? E.g. size of post written to disk? I have "detailed logging" enabled but what it is giving out at the moment is sadly not going to prove anything to you.

Unfortunately Grab-it has no logging feature so I can't even see from that what the error is that the server gives.

[qinglung2]

The only thing so far I could find that remotely sounds like the problem I am experiencing is this:

"The error message "Connection closed unexpectedly by server" means that the server disconnected from your winsock driver without telling your winsock driver that it was going to do so. This is usually caused by an extremely busy server trying to free up resources to handle tasks more efficiently, the server suddenly going offline (because of a lockup, maintenance, power problem, etc.), or phone line noise (which would cause the server to think that you weren't connected anymore)."

I had threads running on both alt.binz and Grab-it when all threads got disconnected (in both programs at the same time), so that's why I believe it must be some error from the server or the connection closing as described above. I know for a fact that the server can get busy, so it is a possibility.

[qinglung2]

Ok, finally got the other error to appear again, though I suspect that it might be a result of an earlier error, i.e. the server disconnecting. To note here: 4 threads were running, downloading post 7, 8, 9 and 10. At 07:49:19 all 4 threads disconnect, Thread 3 gives the msgid error on the retry and Thread 1 and 4 retries correctly. Thread 2 shows that post 8/24 finished, however in the cache folder post 8 has size 8KB, while all the other posts that downloaded correctly have size 646KB. Unfortunately as I said previously, this doesn't prove much to you

[2007/09/14 07:49:06 PM] [T1] Hello
[2007/09/14 07:49:06 PM] [T1] Connecting (IS News Server)
[2007/09/14 07:49:06 PM] [T1] Connected (IS News Server)
[2007/09/14 07:49:06 PM] [T1] Article started 7/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:08 PM] [T2] Hello
[2007/09/14 07:49:08 PM] [T2] Connecting (IS News Server)
[2007/09/14 07:49:08 PM] [T2] Connected (IS News Server)
[2007/09/14 07:49:08 PM] [T2] Article started 8/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:09 PM] [T3] Hello
[2007/09/14 07:49:09 PM] [T3] Connecting (IS News Server)
[2007/09/14 07:49:10 PM] [T3] Connected (IS News Server)
[2007/09/14 07:49:10 PM] [T3] Article started 9/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:11 PM] [T4] Hello
[2007/09/14 07:49:11 PM] [T4] Connecting (IS News Server)
[2007/09/14 07:49:11 PM] [T4] Connected (IS News Server)
[2007/09/14 07:49:11 PM] [T4] Article started 10/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:19 PM] [T2] Article finished 8/24 Lost.S03E07.HDTV.XviD-XOR.r00
[2007/09/14 07:49:19 PM] [T3] Retry 9/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:19 PM] [T3] Asked for <part9of24.jfBsECykcVfgatGap7TM@454354434543jodennnnnn53453.local>
[2007/09/14 07:49:19 PM] [T3] Got reply J®Ç‚>­Ð…‹u©wÊv†ã_‡Ò‰x7f7ÒR\íœ~X×&¡ð¥ËeH½{b`…ïÄE­¹ö!¾«·èÊXÿ^‚ÄÐGÒ0ü0-c;®œü~}†ê«»o}ß“2?Ù½ƒMsê²D!ÿ6åH
[2007/09/14 07:49:19 PM] [T3] Server replied with wrong msgid!!!
[2007/09/14 07:49:19 PM] [T1] Retry 7/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:19 PM] [T1] StatusCode: 0
[2007/09/14 07:49:19 PM] [T4] Retry 10/24 #a.b.mm@EFNet - Req 32967 -Lost.S03 - 268/841 - Lost.S03E07.HDTV.XviD-XOR.r00 yEnc
[2007/09/14 07:49:19 PM] [T4] StatusCode: 0
[2007/09/14 07:49:19 PM] [T2] Connecting (IS News Server)
[2007/09/14 07:49:19 PM] [T3] Connecting (IS News Server)
[2007/09/14 07:49:19 PM] [T1] Connecting (IS News Server)
[2007/09/14 07:49:19 PM] [T4] Connecting (IS News Server)

[qinglung2]

Same problem in the new version  

I'm not the only one with the problem, see posts on this forum:

http://www.newshost.za.net/forum/viewtopic.php?t=211
http://www.newshost.za.net/forum/viewtopic.php?t=226

[Rdl]

I read it but I need more info. I see you have to use 'article cmd'? Did you try without but with retry count>0 What happens then? I need to do a special test version for you to try, so maybe you could come to #alt.binz @EFnet?

[jak3]

I read it but I need more info. I see you have to use 'article cmd'? Did you try without but with retry count>0 What happens then? I need to do a special test version for you to try, so maybe you could come to #alt.binz @EFnet?

Yeah this is a massive problem. It has been preventing me from using alt binz for awhile now. random files end up incomplete. I know of quite a few other people who have the same problem. I'm not talking about corrupt files, which need pars, which is normal, but random incompletes which are downloaded perfectly with other newsreaders.

It would be greatly appreciated if this bug were fixed.

It is quite frustrating not being able to use altbinz, other newsreaders are nowhere near the quality.

[Rdl]

I don't see anyone of you posting observations with the solution I suggested or coming to channell.

[qinglung2]

I read it but I need more info. I see you have to use 'article cmd'? Did you try without but with retry count>0 What happens then? I need to do a special test version for you to try, so maybe you could come to #alt.binz @EFnet?

Without 'article cmd' we get "402 don't have" errors. (I still asked you few months back about the 'article cmd' cause back then we were getting the 402 errors.)

I'll try and connect with IRC but I keep getting disconnected (Software caused connection abort) and Google isn't much helpful in solving it  

[Rdl]

Fixed