http://defect.opensolaris.org/bz/show_bug.cgi?id=14263
--- Comment #1 from Anurag S. Maskey <Anurag.Maskey at Sun.COM> 2010-02-01
18:32:12 UTC ---
There is still an issue here even with my proposed triggering of condition
checks when locations and/or ENMs change state.
Refreshing network location takes about 4-5 seconds. From the SMF logs,
[ Feb 1 17:59:26 Rereading configuration. ]
[ Feb 1 17:59:26 Executing refresh method ("/lib/svc/method/net-loc refresh").
]
activating TestLoc1 location
[ Feb 1 17:59:31 Method "refresh" exited with status 0. ]
And when an NCP is enabled (or NWAM started), there are about 4-5 refreshes of
network/location for a machine with 2 physical interfaces both doing IPv4 and
IPv6 (we refresh network/location every time an interface NCU comes ONLINE/UP).
So, these refreshes of network/location can go on for as much as 30 seconds.
Now, let's say when a location comes online, I force a condition check. When
network/location is refreshed, it tears down everything before setting up the
new location (ipfilters are disabled and then enabled, dns/nis are disabled and
enabled, domainname(1M) is emptied out before getting set again, etc). So,
when nwamd is running its condition checks, it may not find the condition that
it is looking for.
For example, one of the automated tests has condition based on system-domain
provided by domainname(1M). The timing of the network/location refresh and
condition checking is important because the condition checking may not see the
domainname as it is cleared before being set (again).
I do not have an idea about how to workaround this problem or solve it. Any
ideas?
--
Configure bugmail: http://defect.opensolaris.org/bz/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
