cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Santhosh Edukulla <santhosh.eduku...@citrix.com>
Subject RE: Virtual Router reachability
Date Mon, 30 Sep 2013 05:55:51 GMT
Girish,

1. Using timeout will make it to wait for that many units always and again it may not be fool
proof, we may succeed to find ssh daemon on remote machine is up and running  and with this,
we again run it few more times to check again if it is not running, so its not much predictable.


Instead of  waiting for specified time always. 

 a) poll check  to see if ping to ip is working and then verify ssh port is open for connections
on the target, if yes, we are good to go.  This way, we are not waiting for specific time
always.
 b)  There is a telnet lib of which you can use either read_until or expect calls with specific
max timeout ( worst ) with strings like "connected" etc to check ssh port is available for
connections, If we are getting the desired string in the output, then we are ok, or otherwise
you may take a call. This way, we wait for max time only during worst cases.  Check the link
below link for specific examples: 

http://docs.python.org/2/library/telnetlib.html

Regards,
Santhosh
________________________________________
From: Girish Shilamkar [girish@clogeny.com]
Sent: Monday, September 30, 2013 1:38 AM
To: dev@cloudstack.apache.org
Subject: Virtual Router reachability

Hello,

Egress rules tests rely on accessing virtual router VR after creating a network. I have often
seen that VR is not immediately accessible.
A ssh to VR fails, I think it takes a while for the network to come up and VR can be ssh'd.
This happens even though the VR is in "Running"
state.
So I added a delay before trying to ssh VR. I was wondering what is the right amount of delay
here. I did not find a param in global settings
which can be used as wait time.

Please advise.

Regards,
Girish

Mime
View raw message