server priority

[bloggs_&_co]

Well having fiddled with the server setting I would much perfer to see the severs set up with the order of useage in a chain starting with the top
server with highest priority,  article unavailable try next in the chain
and so on, I found that even with all the parts for the article I was d/l`ing
being there that a different program had no problem with just one sever
as top priority, once binz used another server  failed to download the part it stopped for sometime before eventuly getting the part.pausing and unpausing did force some retries but it did seem abit convoluted to get the
parts needed..by the way the queue would be my only criticism at this
point in time the program looks great hope you continue to improve it.

[Rdl]

Hm, I would really like to see log files. Even after all my explaining you still don't get it how it works. It's exactly how you wanted to be
primary server is highest in chain. If primary doesn't have article, program will try with first of the backup server. If that one also doesn't have program will try with 2nd backup and so on.. After trying all primary and backups two things can happen. Either part is missing and file will be incomplete if connection retries is set to 0, or program will try again with primary and backups after defined retry delay. After all this two things can happen. Eventualy part will be available and file will be complete or after defined number of retries (and delays between them) part is skipped and file is incomplete.

When you see that 'part is stopped for sometime' it's cause of set delay between retries. Icon next to file during that time is different and it's obvious that missing parts will be retried for a defined number of retries and with defined pause between them.

[bloggs_&_co]

Well I,m sorry to say but something is not right, the way  the servers are set as backups slows down the download considerable. it has too
go down the chain one server at at time also I currently have five parts just looping round trying the same servers each time, and granted I
do have them set all as prim servers, I have used a second program
with the same servers and these articles have been downloaded ok.
I guess if I leave yours for long enough it might well do the same.
With a bunch of good servers this program as is, will probable work just fine, i have no way to test this. I can`t see it working very well with temp servers that are used to get bits. or to get max download speed.
there is also the possibity that I,m just not making myself clear as to
exactly what i,m trying to do. if I can help clear up ant miss understanding
I would be only to glad to oblige. thanks again.

[bloggs_&_co]

Sorry I should have added that with the servers tagged onto a single
prime server I would have to limit this too 1 connection as my other post
stated all servers have the same connections as the prime server.

At the moment with all as prime what seems to be happening is a
 download takes place once a connection is made but then the connection
is dropped others are tried and eventualy it wil came back and get 1 more
if this is of any help..   sorry if i,m seem a pain in the b*******  

[pieper]

I have a question for better understanding,  What will happen if the same backup server will be used:

Prim1: easynews
Backup1a: quicknet
Backup1b: eweka

Prim2: powernews
Backup2a: quicknet
Backup2b: eweka

Do I have to split/divide the connections between 1a and 2a (same for 1b and 2b) to prevent being disconnected by the server?

[bloggs_&_co]

Well after thinking about the problem of havine only 1 connectiion to some
servers I did split my main server into primes with 1 connection, putting
the backup onto this so we have ser1 back1, ser1,back2  ser1,back3
this solves the problem of the one connection.
but I still had to pause and unpause the connection to finsh the download
as they all went into idle mod.  and pieper not sure if you whom your
question was aimed...sorry :cry:

[pieper]

To Rdl off course,

I got a hunch your retry count is set to 0. Please check this in setup->connection options->retry count. Retry count = 10, retry delay = 5 minutes should be okay. This option is implemented in case a post is not yet complete. It tells ABZ to wait 5 minutes before trying to download again.

The server section seems to be one of the more complex features.  But let me try to explain it in my words of how I would set up the server section. I hope I'm right about this...
I assume you've got a High bandwidth connection (say 10 - 20Mb/s), otherwise there is no need for using multiple primary servers.

Let's define 3 free servers:
Server  max connection
FreeA, 4
FreeB, 4
FreeC, 1
and define 2 payservers:
PayA, 8
payB, 4

I want to download mostly with free servers, so I set them up as primaries:
Primary1: FreeA, 2
Primary2: FreeB. 2
Primary3: FreeC, 1

Now add the backupservers:

Prim1: FreeA, 2
 Backup1.1: FreeB, 1
 Backup1.2: PayA, 3
 Backup1.3: PayB, 2

Prim2: FreeB, 2
 Backup2.1: FreeA, 1
 Backup2.2: PayA, 3
 Backup2.3: PayB, 1

Prim3: FreeC, 1
 Backup3.1: FreeA, 1
 Backup3.2: FreeB, 1
 Backup3.3: PayA, 2
 Backup3.4: PayB, 1

This way there is always multiple downloading from 3 freeservers in parallel, acchieving a very high bandwidth!
If any of the free servers cannot find a particular file, it will search at first on another freeserver, and only at last on one of the 2 payservers.

First assume the answer to my previous question is yes, the connections of one particular server needs to be divided over more than one entries.
This means; all connections must be divided, (assuming the answer to my previous question is yes):
Free A has max connection of 4, so prim1 =2, both backup =1
Free B has max connection of 4, so prim2 =2, both backup =1
Free C can't be used as a backup server since only one connection is allowed.
PayA  has max connection of 8, so bck1 =3, bck2=3 and bck3=2
PayB  has max connection of 4, so bck1 =2, bck2=1 and bck3=1

Personally I believe two payservers is not necessary, one payserver should be enough.

Last but not least, bloggs_&_co you should post your logfile, so Rdl can take a better look into your problem.... :wink:

[pieper]

I just read your other post dealing with the same problem. I assume it is fixed now?

[bloggs_&_co]

Well pieper I had reached the same conclusion as you illistrated on your on your last post, and that was to split myservers with only one conection in order to achive a better download speed and not have the server waiting for a slow connection to finish, i will say I have been using programs such as bnr2 were you can have parallel connection with all available servers and the program will just pickup the next segment in the queue this one is set up in a different way...

Rdl..doing it this way works better so guess i,ll have to play with it some more.  if I understand  how your program works correctly then doing it this way does stop the primary server from carrying on with the download until the backup has finish getting the odd bit thats missing...is so its idle for the length of the slower server?