It's a race condition, we have tuned in 2.7 to 'win' more often, we lose
the race a lot on the first pass.  The DNS issues are unrelated.  We
haven't considered it urgent since the retry serves to rectify it.  The
first pass failing is innocuous.

For info, the issue is the awk daemon we write seems to occasionally take a
surprisingly long time to start and bind the socket.  We probably could
have also fixed it by writing a tiny C daemon, but kept to scripting code
since it's only a few seconds paid for bmcsetup.



From:   Russell Jones <[email protected]>
To:     xCAT Users Mailing list <[email protected]>,
Date:   03/14/2012 10:49 AM
Subject:        [xcat-user] xCAT bmcsetup troubles



Hi all,

Putting together a rack of some Dell c6100 series nodes, and am running
into some strange bmcsetup errors.

Everything appears to be configured properly on the xcat management node,
both the compute and the BMC networks all ping without an issue. When I
tell a node to do a bmcsetup, it displays the following error on the
screen:

Unable to prove root on your IP approves of this request



It then waits for a random period of time, tries again and always succeeds
the second time with no errors being shown.

The management node does not have access to the internet, and as a result
it is also dumping these bind warnings out to /var/log/messages. These only
appear when the compute node is requesting config parameters from the MN:

Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'0/A' (in '.'?): disabling EDNS
Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'./NS' (in '.'?): disabling EDNS
Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'C.ROOT-SERVERS.NET/AAAA' (in '.'?): disabling EDNS
Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'C.ROOT-SERVERS.NET/AAAA' (in '.'?): disabling EDNS
Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'C.ROOT-SERVERS.NET/AAAA' (in '.'?): disabling EDNS
Mar 13 10:46:24 linux7mgt named[6915]: too many timeouts resolving
'L.ROOT-SERVERS.NET/AAAA' (in '.'?): disabling EDNS



Could these issues be a side effect of bind not having access to the
internet to do root DNS lookups? It seems unlikely, but given that it seems
to always succeed a second time, and after succeeding everything works
perfectly, I'm at a loss.


Thanks for any help!

------------------------------------------------------------------------------

Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user

<<inline: graycol.gif>>

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user

Reply via email to