RE: Detecting network failure

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Steve,

I have automated generation of the bootdisk.img, in a way where it takes
params from CGI, builds new ks.cfg based on params, and then stuffs it
into initrd (rdstuff). It's the attached to the system through virtual
media. The process takes about 1 minute to the server booting floppy,
but it's too much time. I was just testing a new piece of equipment that
came in and the server has 4 network ifaces and has feeds to 3 subnets.
I didn't know which iface was eth0 which eth3 I also didn't know what
subnet is associated with which nic. Yes it does sound strange but it
does happen.

Anyways, I have co-workers that are using my build and sometimes they
can't proceed because they can't retrieve the netstg2 image and all
that. So it always comes back to me ;) Some time ago I was speaking to
an Intel hardware vendor about introducing a check-net type of a thing
from their remote management card. Now that the industry is strongly
considering SMASH CLP, this would be akin to Sun's "watch-net"...

So...

I was wondering if there could be a simple thing in anaconda that's
basically - ethtool/mii-tool/tcpdump combo - ok, are we on the subnet?
Right, can we ping the remote server? [if not, display in English -
dude, I'm not sensing the connection, link down ] - right, can we get to
it on port (21, 80, else?) [ if not, dude, your connection is live and
good, yes I'm on the network, but the server is down ] right, is the
image available [ no there's not even a build tree there, pointing to
the right directory ? ] , right - build. Instead there's a "reverse DNS
lookup 5 second"... Well that's cryptic ;) 

Thanks,

Roman Lazarev
Fidelity Investment Management Technology
245 Summer Street V2E
Boston, MA 02210
(617) 563-1173


-----Original Message-----
From: Steven Augart [mailto:saugart@xxxxxxxxxxxxxxxx] 
Sent: Monday, May 15, 2006 6:40 PM
To: Lazarev, Roman
Cc: anaconda-devel-list@xxxxxxxxxx
Subject: Re: Detecting network failure


Lazarev, Roman wrote:

> It also is time consuming to regenerate images every time I want to 
> test another interface.

This comment is intriguing.  What do you mean by "regenerate images 
every time I want to test another interface" ?  Do you mean that you're 
testing different kinds of network interfaces, or is there something 
else going on here?

> Is it possible/is there already a stub that tests network connectivity

> and says - yes, I'm able to communicate - in plain English language ;)

I'm sorry, I don't understand what you  mean by "a stub".  Are you 
hacking on Anaconda and adding features to the Python components?  My 
experience with Anaconda guts is almost entirely with the C loader
stage.

--Steven Augart


[Index of Archives]     [Kickstart]     [Fedora Users]     [Fedora Legacy List]     [Fedora Maintainers]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]
  Powered by Linux