On Wed, Jul 28, 2010 at 01:54:32PM +0400, ????????? ?????? wrote:
Hi,
Now 'scan time' set to 60 sec. The bird load system up to 100% during 10-12 sec every 60 sec.
There is probably something broken in your setup. I have full BGP in kernel routing table and periodic scanning took just a small fraction of second. Perhaps you have some excessive logging that slow it down, or many routes that are unsuccessfully trying to synchronize during each iteration (usually with 'Netlink: No such process' error in logs)?
What circumstances to take into consideration that it is correct to set 'scan time' of protocol kernel?
Periodic scanning is here mostly to catch problems with normal route upload and to catch situations when routes disappear because of some external reason. Probably you can safely set it to several minutes.
If I do not add paths with linux the console, whether I can delete 'scan time' from the protocol kernel?
If you delete it, default (60 sec) will be used. -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."