How do I keep RC5 going for a week?

[DeepBlue]

I just loaded up RC5 on my old 400mhz machine and was just wondering how to make it so it can go for about a week without an Internet connection on its own. Thanks.

 

[vss1980]

In the client there is an option to do this:

Go to configure and then select:
Option 2 (Buffer and Buffer Update options)
Then pick option 14 (Fetch time threshold)
Then type in a time in hours.

Guide:
168 hours in a week

 

[DeepBlue]

Shouldn't I also download a bunch of units to keep it going and what about the flush part? What if it tries to flush prematurely?

 

[vss1980]

That option will make the client download enough units to keep itself busy for the time period you specify. It isn't 100% exact (around 99%) but its the best you can do for a specific time period.

It should only flush by itself when it needs new packets or has a reason to contact a keyserver.
However to stop it flushing until its buffers really are empty you can select an option in the same menu called 10) Additional buffer-level checking, and setting this to option 4. That tells the client to work on a project until its buffers are empty.
Also, if you really want complete connection control you can use option 5) Disable buffer updates from/to a keyserver, and set that to yes. This will mean that the client will not auto-connect to a keyserver and will only update its buffers when you tell it to (bad thing is, if you dont it will crack random blocks). If you do select no, the option to change your server/connection settings disappears but, dont worry, it still works the same as you left them.

 

[Bakwetu]

If it doesn't have an internet connection you have to dl the blocks on a machine that has a connection. This is what I do: shut down client on connected machine, change the dnet.ini file to dl as many blocks as you need, this setting might be appropriate
[rc5]
preferred-blocksize=33
fetch-workunit-threshold=4000

Then lift buff-in, buff-out and cp out of the rc5 folder, restart the client which will now fetch 4000 blocks, after the fetch shut the client down and transfer the new buff-in, buff-out and cp to the non-connected machine, put everything back in the connected machine. It's a bit akward but it works fine for me.