[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[LDM #BIG-900661]: Terminated obsolete upstream



Brice,

> What I meant by mirror and reversed was this -
> 
> Request   FT4  "^WT|^RW|^JS|^LR"
> Request   FT4  "^LR|^JS|^RW|^WT"

Palindromic patterns. Cute!

> That seems to work - no terminations since this morning.  I'll get Richard to 
> try the 'client name' concept and let you know how it works.
> 
> I'm still puzzled by the other obsolete entries with the 'direct-connect' 
> clients.  There in an environment that I don't have administrative access 
> to... yet. So I haven't changed them yet to try anything.  How 'deep' does 
> the 'same server' check go?  I wouldn't think it would break on servers 
> coming out of the same network, but that is the symptom.

The server-check by the LDM is on the IP addresses: requests from the same IP 
address are considered to come from the same downstream host. If your 
'direct-connect' clients are behind a NAT, then that would explain the behavior 
you're seeing.

> I should have access to those clients and be able to restructure them.  If 
> you get tired of this, just point me to the module that does the checks and 
> I'll figure it out.  Currently we still have the source on one of our 
> servers.  In the future though, we are looking at installing via RPM and it 
> doesn't include the LDM source.  May have to keep a copy around in my hip 
> pocket.

The source is (as will likely be for quite a while) on GitHub. The relevant 
file is "ldm_server.c". Search for the string "Reduce".

> Anyway, I'll let you know what else comes up.

Please do.

> Brice

Regards,
Steve Emmerson

Ticket Details
===================
Ticket ID: BIG-900661
Department: Support LDM
Priority: Normal
Status: Closed