On Jun 8, 2016, at 6:48 AM, Steve Clark <steve.clark@xxxxxxxxxxxxx> wrote:
> I must be missing something here, so the system comes up, ip(s) are assigned to the interface, routes, etc then sometime later the switch comes up and you
> ssh in. Never been a problem for me.
Even with static configurations, I’ve had this problem. At least in RHEL6, if the switch doesn’t indicate the interface is up during boot, the ‘network’ service detects the down interface and never starts the network service. full stop. I’ve also seen this happen when the switch has a broadcast storm or some other networking problem and doesn’t become active for more than a minute after boot. Often I’ll have to add a line to the ifcfg-* script to have it just sleep for 60 seconds before even trying to activate the interface, when I know the system is on a switch that takes a long time to perform its splay tree calculation. (Many of my systems are on networks I have no control over, so I have to just work around problems like this.)
I’ve always used NM in RHEL7 so I’m not aware if systemd is smart about dynamic interface activation of the ‘network’ service. NM in RHEL7 is so much better than in RHEL6 so I haven’t really needed anything else.
--
Jonathan Billings <billings@xxxxxxxxxx>
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
https://lists.centos.org/mailman/listinfo/centos
