Diffrent computers using shared folder

[mcriley]

Hi,

I quite a few computers that I manage and would like to run SoB on them.

However, they are not really used regularly and some get used more than others and different computers are used at diffrent times. Some might be unused for weeks and some might be on more or less 24/7 for some time (not the same all the time though)

I was thinking, instead of running SoB on them all separalty (some machines would probably take forever to complete a test), would it be possible to set up one machine, share a folder with the program from it and have the rest of the computers use that folder? Or will it mess everything up?

I am probably trying tomorrow but thought it would be nice with some input if anyone understands what I mean and has a clue if it is a good or a lousy idea.

Regards

[Joh14vers6]

Hi,

I quite a few computers that I manage and would like to run SoB on them.

However, they are not really used regularly and some get used more than others and different computers are used at diffrent times. Some might be unused for weeks and some might be on more or less 24/7 for some time (not the same all the time though)

I was thinking, instead of running SoB on them all separalty (some machines would probably take forever to complete a test), would it be possible to set up one machine, share a folder with the program from it and have the rest of the computers use that folder? Or will it mess everything up?

I am probably trying tomorrow but thought it would be nice with some input if anyone understands what I mean and has a clue if it is a good or a lousy idea.

Regards
In Windows it is not possible due that the k/n-pair is written in the register. I think with Linux it is possible.

[umccullough]

In Windows it is not possible due that the k/n-pair is written in the register. I think with Linux it is possible.

It is possible to move a test from one windows machine to another by simply editing the registry of the "target" computer to use the same k/n pair before starting the client. At this point, 2 machines cannot simultaneously work on the same k/n pair, however.

Not too long ago, I moved a bunch of tests off my slower machines (P3s, etc.) onto some faster P4's in order to complete them and reduce my pending test count. This was just a matter of setting the registry k/n keys, setting the cache key to '2' and making sure the test file was in the directory. Once the client is started, it will pick up where the other machine left off.

One key is you have to remember to have the client stop when the proth test is complete if you want to manually set up the next test... it takes some attention.

Read the post on recovering a "lost proth test" - it pretty much has all the info you need.

I am curious, however, if multiple machines running sob from the same directory will successfully share the same log file?

[mcriley]

In Windows it is not possible due that the k/n-pair is written in the register. I think with Linux it is possible.

Ok, too bad. It would be nice if it was possible.

Hopefully even the slow machines that are not used much will be able to complete some tests nevertheless.

Regards

[Keroberts1]

you shoudl check into sieving. That would be much easier to distribute and alot of slow machines is just as good as one fast one in sieveing.

[Joh14vers6]

It is possible to move a test from one windows machine to another by simply editing the registry of the "target" computer to use the same k/n pair before starting the client. At this point, 2 machines cannot simultaneously work on the same k/n pair, however.

Not too long ago, I moved a bunch of tests off my slower machines (P3s, etc.) onto some faster P4's in order to complete them and reduce my pending test count. This was just a matter of setting the registry k/n keys, setting the cache key to '2' and making sure the test file was in the directory. Once the client is started, it will pick up where the other machine left off.

One key is you have to remember to have the client stop when the proth test is complete if you want to manually set up the next test... it takes some attention.

Read the post on recovering a "lost proth test" - it pretty much has all the info you need.

I am curious, however, if multiple machines running sob from the same directory will successfully share the same log file?
The TS meant to do this automaticaly, not manuely each time.
If someone can made a program SBPool what would work a like SBQueue and that set the registry to [HKEY_LOCAL_MACHINE\SOFTWARE\LhDn\sob\cache]"cache"=dword:00000000 to force a new test at shutdown. The program should remember which tests are not finished at which tests are in use.

The logfilename can be changed in the clientconfiguration for each client.

[Ken_g6[TA]]

If someone can made a program SBPool what would work a like SBQueue and that set the registry to [HKEY_LOCAL_MACHINE\SOFTWARE\LhDn\sob\cache]"cache"=dword:00000000 to force a new test at shutdown.I had given some thought to this.
The program should remember which tests are not finished at which tests are in use.There's the rub. Finding whether or not a client is running is next to impossible. :(

You might be able to do something with SBPrp (http://home.att.net/~k.brazier/programs/java/sbprp/), manually inserting your tests, and setting the timeout time to some low but not too-low value. (The timeout value is in the Java code, BTW.) The current timeout is 1 day, I believe.