Pavlin Radoslavov wrote: >> However, you can imagine that at any time an interface may leave, and if it >> leaves before Xorp notices, then you can get those crashes. I'm all for >> having >> errors in the logs, but allowing an external event to crash the router seems >> like a bad idea. >> > > Absolutely agree, an external event should not crash the router. > Again, the assert is there to tell the developer there is an unusual > problem that needs to be investigated. > It seems to take a while to propagate information through xorp, so it is possible for the code to not yet have discovered the interface is gone, and yet try the system call to get the interface name/index.
There is absolutely no way you can be *certain* an interface is in the linux kernel when you make that call, so you *have* to be able to handle failure cases if you want your app to be robust. > If the tool can be used to easily reproduce the problem, then yes > please send me a copy of it with simple instructions how to use it > to trigger the problem. > I'm not sure it will reproduce this particular issue, but it should help you with other testing, especially with virtual routers and such. I'll send you information off-list on how to get it set up. Thanks, Ben > Thanks, > Pavlin > -- Ben Greear <[EMAIL PROTECTED]> Candela Technologies Inc http://www.candelatech.com _______________________________________________ Xorp-hackers mailing list [email protected] http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers
