Problem with distributed network

Discussion and support for the CUDA Multiforcers (Windows and Linux)
  • Ads

Problem with distributed network

Postby schnaps » Thu Jul 26, 2012 5:07 pm

Hi Bitweasil!

I'm here to report an issue with the server mode. As it seems, if for some reason a client goes down (i.e. connection problems), the server keeps announcing that the client is still there, although we can see there is no speed variations, which happens with connected clients. I'm not sure if this is the only case where this happens, I need some further testing. As I'm thinking, it should be possible to implement a solution where the server checks client connectivity from time to time, and if the client doesn't answer within a period of time, it would be removed from the list and it's current WU would be redirected to some other host (or marked as "undone"). Can this be done?

As a side request, is there a way to compile the source on 32-bit Linux systems? Do you know of some guides that could help me do this? Since I'm thinking on using a 32-bit system to run in serveronly mode, limitations with rainbow tables shouldn't be a problem.

Thanks in advance!
Cheers!
schnaps
 
Posts: 7
Joined: Tue Jul 24, 2012 9:13 pm

Re: Problem with distributed network

Postby Bitweasil » Thu Jul 26, 2012 6:09 pm

Aware of this issue and will be implementing timeouts. Thanks for reporting it!

As for 32-bit, there should be a flag in Cmake for 32-bit.
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm

Re: Problem with distributed network

Postby schnaps » Thu Jul 26, 2012 6:23 pm

Bitweasil wrote:Aware of this issue and will be implementing timeouts. Thanks for reporting it!

As for 32-bit, there should be a flag in Cmake for 32-bit.

Thanks for the fast reply. So setting the Cmake flag is enough? No need to alter anything else?
schnaps
 
Posts: 7
Joined: Tue Jul 24, 2012 9:13 pm

Re: Problem with distributed network

Postby Bitweasil » Thu Jul 26, 2012 8:35 pm

schnaps wrote:Thanks for the fast reply. So setting the Cmake flag is enough? No need to alter anything else?


If it's implemented properly... :)

(I didn't actually do the build system, I just maintain it - so if some edge stuff like 32-bit is broken, I can file a bug with the maintainer of it)
Bitweasil
Site Admin
 
Posts: 912
Joined: Tue Jan 20, 2009 4:26 pm


Return to CUDA Multiforcers

Who is online

Users browsing this forum: No registered users and 2 guests

cron