>>>>> "Aleksi" == Aleksi Suhonen <Aleksi.Suhonen@xxxxxx> writes: >>>>> "Kalt" == Christophe Kalt <kalt@xxxxxxxxxxx> Writes: Kalt> These problems occur when there are several simultaneous Kalt> server connections being established. The result is a server Kalt> collision and it isn't always dealt with correctly. Kalt> A way to avoid this is to tune the auto connections, and be Kalt> careful to avoid having everything auto connect to everything Kalt> else. This is a little ambiguous -- I'll assume you mean that it's bad to have A configured to auto-connect to B as well as B configured to auto-connect to A. Aleksi> Yes, on larger networks the standard is that the minimum Aleksi> server autoconnect period is two to five minutes. This Aleksi> usually avoids the problem. On smaller networks you might Aleksi> want to have smaller autoconnect periods but in that case I Aleksi> would suggest that you make sure that the autoconnect Aleksi> interval of each C:line on the whole network is a different Aleksi> prime number. It's also convention to arbritarily define a "direction" of server connections, and only auto-connect to servers "upstream" while relying on "downstream" servers to connect to you. If I'm understanding Kalt correctly, then it's this that matters (due to what sounds like a race condition), not the auto-connect frequency (which only masks the problem). It's also why you've got to be careful when doing a manual /connect -- an auto-connect might be underway. - Andrew -- #!/usr/bin/env python print(lambda s:s+"("+`s`+")")\ ('#!/usr/bin/env python\012print(lambda s:s+"("+`s`+")")\\\012') print(lambda x:x%`x`)('print(lambda x:x%%`x`)(%s)')
Attachment:
pgp00009.pgp
Description: PGP signature