Newsbin 6 error problem. Download stops suddenly

[polakis]

I had newsbin pro 5.30 and all worked fine. I installed 6.10, i made all the properly registration, servers etc.

Then i am starting to download files from usenet. Suddenly dowload stop completely and doesnt restart. Connections are fine and in connected status... Whats wrong?

Please help...

[itimpi]

Does it say anything in the Logging tab?

[Quade]

Are any of the files paused? How about the main pause button?

[Ace]

I just had this or a very similar problem, and it's a real showstopper for version 6.10 for me. I'll keep it up for a little while in case I get a quick reply asking me to check something.

The trigger for me was an internet outage. I know because I use NetWorx to monitor my internet bandwidth and suddenly my upload and download bandwidth went to zero. Here's the screenshot:


I get maybe up to 5 of those a day, they usually auto-reconnect after 3-5 minutes. It's a very rare day that I don't get any such outages. The green is download bandwidth, and it's short at first where just the P2P app is running, but the green bars get much taller when I start the Newsbin download. Then you can see both upload and download bandwidth go to zero. Either that was the trigger or it's a huge coincidence.

Once the internet connection reconnected, my p2p application automatically resumed (as you can see in the graph), but nothing happened with Newsbin, and I mean nothing. The log file just stopped logging, at least there are no messages in the logging tab after the internet connection died (Edit: At least there weren't for 2 hours, see update below).

Versions 4 and 5 dealt with this just fine. They would keep trying to download, and I'd get connection failure messages in the logging tab, and they would fail the parts after retrying so many times. Then when the connection came back, if the outage was long enough, I'd have some failed files or parts, but it would resume. Then I would just go to my failed files list, and add the failed files back to the download list and everything was fine.

I looked at newsbinpro64.exe in my task manager (using the 64 bit version on XP64 if that makes a difference) and it looks like it's still "running" as every once in a while I see the CPU usage column flicker from 0 to 1%, and it's using 133 megs of memory.

It's kind of like it just died, but it's definitely not locked up because I can type a search term at the top and it will parse the headers to the search term.

Anyhow, I don't think I can switch to version 6 until this problem is fixed. But it was doing well in my testing until this happened.

Whether you can duplicate this by unplugging a network cable or not, I don't know, I haven't tried.

I suspect restarting will probably get it going again just fine, but I don't want to have to restart. I prefer the way version 5 works where no restart is necessary.

I saved screenshots of the Newsbin screen and the log file, but the log file is useless since it doesn't show anything after it stopped working. Anything else I need to check before I close it down and restart?

I did check the main pause button, and it looked unpaused (and the symptom is very much like pause), but just to make sure, I clicked it, watched it turn yellow, and clicked it again to make sure it's unpaused (the yellow disappeared). It's like it paused and won't unpause but the pause button isn't on and none of the parts look like they are paused, I see green everywhere on the download tab, like it's still downloading.
-----------

Upadate: I left it running, and almost exactly 2 hours after it stopped, I got a popup window of critical error messages which also appeared in the log. Here's what the log looks like:
[19:18:45] HIGH NNTPSocket - ARTICLE <48a3071c$0$33699$892e0abb@auth.newsreader.octanews.com>
[19:18:45] HIGH NNTPSocket - 220 0 <48a30702$0$33699$892e0abb@auth.newsreader.octanews.com>
[19:18:45] HIGH NNTPSocket - 220 0 <48a3070f$0$33699$892e0abb@auth.newsreader.octanews.com>
[19:18:45] HIGH NNTPSocket - 220 0 <48a3071c$0$33699$892e0abb@auth.newsreader.octanews.com>
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:51] 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.
[21:18:51] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[21:18:52] HIGH NNTPSocket - MODE READER
[21:18:52] HIGH NNTPSocket - MODE READER
[21:18:52] HIGH NNTPSocket - MODE READER
[21:18:52] HIGH NNTPSocket - MODE READER

19:18:45 corresponds to the time where the internet connection was lost. The connection was restored automatically and the P2P app resumed at about 19:22. And Newsbin went into the "twilight zone" for 2 hours, and started running again at 21:18! Very strange! I can't think of anything I did to get it started again, it seems like it just decided to restart on its own, after taking a nap for 2 hours.

On the plus side, it didn't actually fail any files, it picked up right where it left off!

[tl]

19:18:45 corresponds to the time where the internet connection was lost. The connection was restored automatically and the P2P app resumed at about 19:22. And Newsbin went into the "twilight zone" for 2 hours, and started running again at 21:18! Very strange! I can't think of anything I did to get it started again, it seems like it just decided to restart on its own, after taking a nap for 2 hours.

On the plus side, it didn't actually fail any files, it picked up right where it left off!

Did you have a lot of files in the Download List that had been downloaded but was still listed? Obvious ways to get that is that the files were waiting for assembly (missing pieces perhaps and not old enough to switch to repair mode) or that you have autounrar disabled, but there's likely other ways too.

The reason I ask is that I've once seen NB 6.10 (probably an RC version) do nothing visible for extended periods, this was after I paused the download and then resumed it. There was no obvious log messages and very little CPU/network/disk utilization for an extended period, then it suddenly started downloading the remaining item!

It wasn't "2 hours", only 15-30 minutes or so, but if it scales with the amount of things it has to check in the Download List it could potentially scale up to those levels...

My guess is that it was doing some kind of checking on these already downloaded items, but very inefficiently (little CPU/network/disk activity). If so Quade might not have noticed because I think the developers all run NB in "AutoUNRAR" mode where this is much less likely to happen (because files gets removed automatically from the Download List so it's much smaller).

I've only seen this once but if you can confirm that you had a lot of already downloaded items in your Download List we may have enough for the NB developers to be able to locate and fix it. In my case the "Remove from download list if OK" option that Quade has said may be coming may well make this a non-problem.

[Quade]

I'm going to try to simulate this by pulling the plug on my router.

[Ace]

Did you have a lot of files in the Download List that had been downloaded but was still listed?

Thanks for the questions/suggestions om things to look at.

I only had one item from the compact view in the download list.

Expanded, it consisted of 49 avi parts, and 12 par files, so there were a total of 61 items in the download list. It was in the middle of downloading part 26 of 49 when I lost my internet connection. That part 26 looks like it has about 40 chunks, of which 10 had successfully downloaded and it was downloading chunks 11-18 when I lost my internet connection. That's exactly how it stayed for 2 hours.

Now regarding parts 1-25 which had already fully downloaded, I'm not sure how to answer your question because versions 5 and 6 of Newsbin work differently. Had I been using version 5, the parts 1-25 would have disappeared from the download list when they completed. In version 6, they just stayed there, with all the chunks showing as green. So yes I had 25 parts still on the download list, each showing all green chunks, no missing chunks or incompletes, and they don't disappear like they used to with version 5, are they supposed to? I thought that was normal for version 6 that they'd stay on the download list.

I'm using autopar and autounrar. I just downloaded another file and nothing disappeared from the download list until I had everything, then it assembled just fine.

[Ace]

I'm going to try to simulate this by pulling the plug on my router.

I just replicated it by pulling the power plug on my cable modem in the middle of a download. I waited about 2 minutes and plugged it back in. Same problem.

It can't be caused by a long list of items on the download list because I replicated the problem during the download of the very first item! Same symptoms except I don't know if it will take 2 hours to start working again. It's not working 5 minutes after reconnecting though,and it should be. I didn't feel like waiting 2 hours so I just restarted Newsbin and it works fine after a restart; it automatically resumed the download that became "paused without using the pause button".

Edit to add: I also did the same test on version 5.59. The very instant my internet reconnected, I got the error messages about the NNTP server timeout and the download resumed just fine.

Let me know what you find.

[Quade]

By default, there should be a 5 minute delay between pulling the cable and connection timeouts in Newsbin. Depending on failure mode, there's no way for an application to know that no more data is coming. For example, if I have my router working, and Newsbin downloading and I then pull the ethernet cable that feeds my router (not unplug the router). The data just stops but, there are no errors at the TCP level. That's when connection timeouts kick in. It sounds like you have a large connection timeout for some reason (maybe some setting in Newsbin)


//
// You can change the timeout interval here
//
HKEY_CURRENT_USER\Software\DJI Interprises\Newsbin50\Settings\RecieveTimeout

I wouldn't set it less than 60. Is this SSL? Would be interesting to see if the same thing happens when you're not SSL.

[Ace]

I don't use SSL.

I went through what I thought to be the relevant settings in Newsbin to see if anything looked odd..On the Network tab,the first box is checked for show NNTP commands, but none of the other boxes are checked, so the timeouts are grayed out. SSLV2 is also unchecked.

I looked at the location you specified in my registry. There is no entry at that location for "receive timeout", I have entries of Bytes (has a big number), BytesHigh (0), InstallDir, Show MOTD is enabled and SSL is disabled.

Are you saying I need to manually add a registry entry to get Version 6.10 to work, even though version 5.59 works without adding any registry entries?

I was also curious about the registry path including Newsbin 50. I have newsbin 5.59 and 6.10 installed in separate program directories and they have separate data directories.

If the receive timeout settings for version 6.10 are in the Newsbin50 path, where would I find the registry settings for version 5.59? Since version 5.59 works fine and version 6.10 doesn't, would comparing them provide any clues? I've compared settings in the program and didn't find anything but I'm not sure where to look to compare registry entries.

[Quade]

Yeah. Have to manually add it. All versions of Newsbin use the same registry area. It doesn't store a whole lot there. The lion's share of the configuration is in the NBI file.

[Ace]

Actually version 4 uses a separate registry area called "Newsbin Pro".

OK, I closed down Newsbin, edited the registry by adding new string value named receive timeout with value 100. Is that 100 seconds or what? You said not to go below 60 so I figured 100 wouldn't be cutting it too close.

I restarted Newsbin 6.10, started a download, and repeated the test of unplugging the cable modem power cable for 2 minutes. My connection came back about a minute after I plugged the power cable back in,but Newsbin 6.10 never resumed. It's been 15 minutes now. Version 5.59 would definitely be running as soon as I got my connection back (which was 12 minutes ago) had I been using that version.

Do I need to reboot the PC for that registry setting to take effect?

Edit to answer my own question: Yes, a reboot was needed. Now I'm seeing error messages after the internet outage, as I should. My testing has been hampered by the fact that in addition to the simulated internet outages, I'm having some real internet outages too, so it's not always reconnecting when I plug my cable modem back in, like it did this morning. I'll test it more with this setting, but it looks a little better so far, but here's the problem:

Only one connection out of eight resumed. The other 7 connections are still stalled. I never had this problem with version 5.59. Here's a screenshot:

The red failed chunk with the seven stalled connections before it is where the download was when I disconnected the internet. As that screenshot hopefully shows, the one connection has been working all by itself since it resumed, so it's really slow compared to the normal 8 connections I use.

When I tested this with version 5.59, all 8 connections immediately resumed when I regained the internet connection. So this added registry entry is an improvement, but it's not working as well as v5.59 yet.

[Ace]

I tried it again and it did the same thing, so it's reproducible. Do these log entries help at all in determining why 7 of the 8 connections remain stuck for two hours after my internet connection disconnects, then reconnects? I get these errors right after reconnecting, and have 7 stuck connections, with one connection working.

[08:55:12] ERROR Socket - Error: WSARecv Socket Error Code: 10065 A socket operation was attempted to an unreachable host.
[08:55:12] ERROR NNTPSocket - SendCommand Worker - Server:news.giganews.comSendCommand - failed to Read Response From server: Socket - Read From Server Failed
[08:55:12] ERROR NNTPSocket - Socket HandleError - Server: news.giganews.com SendCommand - Server Disabled Service on connection:
[08:55:12] ERROR Socket - Error: WSASend Socket Error Code: 10038 An operation was attempted on something that is not a socket.
[08:55:12] ERROR NNTPSocket - SendCommand Worker - Server:news.giganews.comSendCommand - failed to send command to server: Unspecified Network Error
[08:55:12] ERROR NNTPSocket - Socket HandleError - Server: news.giganews.com SendCommand - Server Disabled Service on connection:
[08:55:24] ERROR SocketAddress - Error: Unable to Convert Server Name to Internet Address: news.giganews.com11001

Is that 11001 supposed to be there in the news.giganews.com11001 error? That looks odd to me but I admit I'm not used to reading these error logs because in all the time I've use Newsbin, I haven't had many unexpected errors where I needed to look at the log.

As before, in version 6.1 the stuck connections stayed stuck for about 2 hours, then they became unstuck.

When I run version 5.59 I don't see any delay at all after the internet reconnects, all 8 connections resume working.

Also, I only installed this XP64 install a week ago on a newly formatted hard drive, so it's a clean install. I also kept a log of everything I've done since installing it, including all tweaks to the operating system. I don't see any changes I made that should affect this or cause a 2 hour freeze of the connections, and moreover I'm not sure how or why it only occurs on version 6.10 version and not version 5.59.

I'm willing to keep running version 5.59 if there's no way to get 6.10 or 6.11 or a subsequent version 6 release to solve this, but I like version 6 so I was hoping I could use it.

[astrax]

Have been experiencing similar problems the last few days. Download speed goes down to 0 and nothing happens after that. There are no error messages or anything, nothing is paused and my internet connection is still active... Only a restart of the program will get the downloads going again.

[Quade]

I pulled the cable from the 'net side of my router. About 120 seconds later, these started up

[17:06:08] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[17:06:08] ERROR NNTPSocket - SendCommand Worker - Server:news.giganews.comSendCommand - failed to Read Response From server: Socket - Read or Write Timeout
[17:06:08] ERROR NNTPSocket - Socket HandleError - Server: news.giganews.com SendCommand - Server Disabled Service on connection:
[17:06:08] ERROR NNTPSocket - SendCommand Worker - Server:news.giganews.comSendCommand - failed to send command to server: Unspecified Network Error
[17:06:08] ERROR NNTPSocket - Socket HandleError - Server: news.giganews.com SendCommand - Server Disabled Service on connection:
[17:06:08] HIGH Download Engine news.giganews.com - INT_OpenItem -DL_DI - Failed to retrieve Article from Server::news.giganews.com
[17:06:08] DEBUG Download Engine news.giganews.com - NNTP_CloseChunk - closing a dirty connection
[17:06:08] ERROR - SSL Read Failed: 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.

[17:06:08] ERROR NNTPSocket - Server:news.giganews.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[17:06:08] DEBUG Download Engine news.giganews.com - NNTP_CloseChunk - closing a dirty connection
[17:06:08] ERROR - SSL Read Failed: 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.

15 of these started

[17:06:16] ERROR NNTPSocket - Server:ssl.astraweb.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[17:06:16] DEBUG Download Engine news.giganews.com - NNTP_CloseChunk - closing a dirty connection
[17:06:16] ERROR - SSL Read Failed: 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.

More from astra

[17:06:16] ERROR NNTPSocket - Server:ssl.astraweb.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[17:06:16] DEBUG Download Engine news.giganews.com - NNTP_CloseChunk - closing a dirty connection
[17:06:16] ERROR NNTPSocket - Server:ssl.astraweb.com NNTPServer: Timeout - Waiting for Server to send Data: Socket - Read or Write Timeout
[17:06:16] DEBUG Download Engine news.giganews.com - NNTP_CloseChunk - closing a dirty connection

10-15 of these per server 1 per connection

[17:06:20] ERROR SocketAddress - Error: Unable to Convert Server Name to Internet Address: news.giganews.com11001
[17:06:20] ERROR SocketAddress - Error: Unable to Convert Server Name to Internet Address: ssl.astraweb.com11001

I plugged my cable back in and Newsbin re-connected almost immediately.

[17:06:41] HIGH InterSocket - SSL Zlib Compression Added
[17:06:41] HIGH InterSocket - SSL Zlib Compression Added
[17:06:41] LOW InterSocket - SSL Connection Server: news.giganews.com Stats: AES256-SHA 256 TLSv1/SSLv3
[17:06:41] HIGH InterSocket - SSL Zlib Compression Added
[17:06:41] LOW InterSocket - SSL Connection Server: news.giganews.com Stats: AES256-SHA 256 TLSv1/SSLv3
[17:06:41] LOW InterSocket - SSL Connection Server: news.giganews.com Stats: AES256-SHA 256 TLSv1/SSLv3
[17:06:41] LOW InterSocket - SSL Connection Server: news.giganews.com Stats: AES256-SHA 256 TLSv1/SSLv3
[17:06:41] LOW InterSocket - SSL Connection Server: news.giganews.com Stats: AES256-SHA 256 TLSv1/SSLv3

This happens because I was close to my limit when I pulled the cable so, I'll be over the connection limit for a few minutes.

[17:06:41] ERROR NNTPSocket - NNTP Perform Auth - news.giganews.com-The server rejected your login - contact your news service's support department. It's not a Problem with Newsbin : 481 (remote) (aucl:gn;gn210951) exceeded maximum number of connections per user

At this point. I'm not sure why you're seeing what you're seeing. You can see I'm using SSL here. Are you using SSL? I doubt SSL matters one way or the other.


I don't think I've ever seen this error before.

[08:55:12] ERROR Socket - Error: WSARecv Socket Error Code: 10065 A socket operation was attempted to an unreachable host.

Is there something special about your setup? Are you dual homed or something?

[zwolanski]

The same problem with NB Pro 6.11 Bd 1342. After service interruption the NB fails to resume download after service restoration. The outage lasted less than a minute, NB did not resume for 1 hour. Instead of waiting longer, I restarted the program. It started downloading everything from the begining creating copies of already downloaded posts ! Not good.

[Quade]

Same symptoms Ace reported. I wasn't able to reproduce it.

When 6.2B1 on comes out. Try it. If you get the same symptoms, I'll add extra logging so, we can figure out what's happening.

[Ace]

Same symptoms Ace reported. I wasn't able to reproduce it.

When 6.2B1 on comes out. Try it. If you get the same symptoms, I'll add extra logging so, we can figure out what's happening.

Sounds great, I'll try it when it comes out.

At this point. I'm not sure why you're seeing what you're seeing. You can see I'm using SSL here. Are you using SSL? I doubt SSL matters one way or the other.

You asked me that before and I answered but I guess you missed it, I'm not using SSL.

[08:55:12] ERROR Socket - Error: WSARecv Socket Error Code: 10065 A socket operation was attempted to an unreachable host.

Is there something special about your setup? Are you dual homed or something?

My friend is dual horned so he can isolate part of his network from his kids, so I'm familiar with that setup he's using, but I'm not using anything like that.

You may be right about that error being unusual, I ran a google search of the entire web, and got one result on that, and the result was this thread! I then widened the search and found this but it wasn't too helpful: http://www.sockets.com/err_lst1.htm#WSAENETUNREACH

I can't think of anything that unusual about my setup except I may be the only person on the planet with the registry modification you asked me to make, but that's probably not it, right? Obviously there's some difference with yours since you're not seeing it (Thanks for looking). My motherboard has 2 network interfaces, but I'm not using the second one, and though the more typical motherboard has only 1 network connection, I don't think mine is that unusual.

Anyhow, I look forward to trying 6.2B1. Thanks!

[Ace]

Quade, I don't know how you managed to fix this problem since you said you couldn't duplicate it, but you did fix it!

I just started running Newsbin pro 6.20B1 build 1423 64 bit about 15 minutes ago.

First, loading the headers spool happened almost instantly!

Then I tested the problem this thread is about, by disconnecting the cable modem power, and reconnecting it.

I didn't have the problem of stuck connections like I had with version 6.1, but I did have one failed chunk (turned red as in my previous screenshot) that I suspected was due to the registry entry you asked me to add on "receive timeout".

So, I deleted that registry entry, and tried again, twice, and now I don't even get that failed chunk when my service is temporarily interrupted! Every single chunk downloaded and the connections resumed normally as in version 5.59. Yay! So the latest beta actually works better without that registry modification. I haven't had time to test much else in 15 minutes, but I thought I'd let you know you fixed it!

Thank you, thank you, thank you!

[Quade]

Man, that makes me happy too. I hate these weird "happens to you, doesn't happen to me" problems.

[r2d3p0]

I am getting this same error
[18:57:43] ERROR SocketAddress - Error: Unable to Convert Server Name to Internet Address: us-secure.newsdemon.com11001

Everytime Newsbin starts
My downloads also stop suddenly at time with no errors reported

This error was also happening on the non-ssl servers

I am useing Newsdemon

[Quade]

Error: Unable to Convert Server Name to Internet Address: us-secure.newsdemon.com11001

This is a DNS error. It's coming from windows.

> server 8.8.8.8
Default Server: google-public-dns-a.google.com
Address: 8.8.8.8

> us-secure.newsdemon.com
Server: google-public-dns-a.google.com
Address: 8.8.8.8

Non-authoritative answer:
Name: iad.sslusenet.com
Addresses: 69.16.185.248 69.16.185.251
Aliases: us-secure.newsdemon.com