using primary and backup servers

[jarthel]

my ISP has a caching usenet server and this allows me to have 6 connections. Surprisingly I get faster speeds on this caching server (no idea on the peer) than giganews.

As such, I setup my ISP server as primary and use giganews as backup.

But somehow, most the files I get are corrupted and giganews was never used to download the missing bits. As a test, I downloaded 2 file using my ISP as primary and giganews as backup. Both files was incomplete according to the built-in par checker and quickpar.

I then download the same files using giganews. both files came up as complete.

any ideas? maybe a bug somewhere?

[Rdl]

my ISP has a caching usenet server and this allows me to have 6 connections. Surprisingly I get faster speeds on this caching server (no idea on the peer) than giganews.

As such, I setup my ISP server as primary and use giganews as backup.

But somehow, most the files I get are corrupted and giganews was never used to download the missing bits. As a test, I downloaded 2 file using my ISP as primary and giganews as backup. Both files was incomplete according to the built-in par checker and quickpar.

I then download the same files using giganews. both files came up as complete.

any ideas? maybe a bug somewhere?

Incomplete <> missing

Backup servers are used for missing articles. I presume that article is corrupted on your primary server.

[davidq666]

so a file marked as incomplete could either have fileparts(articles) missing or not missing but broken? is that about it? if so is there a way to tell which is which?

i see many advantages in backup server not just fill in missing files but fill in missing fileparts.... maybe the backup server can monitor in the primary in real time and then fill in missing file parts in real time , just like anti virus monitors computer operation in real time .

But this is exactly what Alt.Binz already does ... articles = "fileparts" in the sense that you mean it.  On the PAR2 tab, the term 'blocks' is used.  Read your log, and you should see Alt.Binz downloading parts of a file - e.g. 'Article started 1/40 my1337file.rar' - and if one of those parts is missing, it will check your backup server for it when the next thread is available, i.e. as close to "real time" as Alt.Binz gets.

[Jesus]

Backup servers are used for missing articles. I presume that article is corrupted on your primary server.

Most modern Usenet downloader programs will try a backup server if there is a yEnc CRC mismatch.  In some programs that's not the default option, and it has to be turned on.

[jarthel]

I read Hecks reply on another thread and I thought backup servers are used for incomplete/corrupted files?

so a file marked as incomplete could either have fileparts(articles) missing or not missing but broken? is that about it? if so is there a way to tell which is which?

i see many advantages in backup server not just fill in missing files but fill in missing fileparts.... maybe the backup server can monitor in the primary in real time and then fill in missing file parts in real time , just like anti virus monitors computer operation in real time .

But this is exactly what Alt.Binz already does ... articles = "fileparts" in the sense that you mean it.  On the PAR2 tab, the term 'blocks' is used.  Read your log, and you should see Alt.Binz downloading parts of a file - e.g. 'Article started 1/40 my1337file.rar' - and if one of those parts is missing, it will check your backup server for it when the next thread is available, i.e. as close to "real time" as Alt.Binz gets.

[Jesus]

Corrupted Usenet articles are rare now in 2008, very unlike 2005-07.  I wonder if something else is going on.  It must be frustrating to troubleshoot something like this without being able to examine a detailed NNTP log.  I see topics like this all the time, someone attempting to troubleshoot a downloading issue, but no NNTP log is available to help provide answers.

[Hecks]

I read Hecks reply on another thread and I thought backup servers are used for incomplete/corrupted files?

No, there's a difference between incompletes ( = article can't be found on server) and corrupted artcles (ones that fail the CRC check).  Your log will always tell you which.