[ale] Rackspace Cloud Server Issue

[Chris Fowler]

I experienced an issue with a cloud "server" at RS this morning that 
took me over 2 hours to solve.  Luckily the system was not critical, but 
I'm sending this email because I thought it was a bit odd how it was solved.


Early this morning Thunderbird alerted me that it failed to connect to 
imap.k4fh.com.  I verified that system was down with failed ping.   
Logging into the cloud control panel showed it as up so I did a hard 
reboot.  After the reboot there was no change so I tried to connect to 
the console.  That was not working either so I started a chat session.

My guess is that I was chatting with someone outside the US that had 
access to second level at RS.  They started working on the problem and 
told me that the host system had to be restarted. During this period I 
checked status.rackspace.com and saw now down reports.  I did receive 
and email upon the reboot that RS was working on the system.  After the 
reboot I was still unable to ping, but this time I was able to connect 
to the console.  Logging in I was able to ping RS's private subnet on 
eth1, but nothing on the public at eth0.  The person I was chatting with 
assumed everything was back up.  I said different.  After about 10 more 
minutes I was able to ping out and I was able to ping the system.


A few things concern me.

1.  I became their monitoring system that told them they had a problem 
and they chose to reset the host.
2.  After reset the problem was not resolved.  I'm guessing a failed 
routing issue on their end and the host was fine.
3.  During this period the person I was chatting with assumed the 
problem was on my end and after the host was up also assummed my guest 
was functional.
4.  Resetting a VM host is the last thing one would want to do. All the 
guests are restarted as well.

Chat Transcript
https://dl.dropboxusercontent.com/u/2486826/RSCHAT.pdf

This system was a personal guest.  I have a business account with RS 
where I have many guests.  I've also bee recommending them to customers 
that want to move our stuff off real hardware and onto cloud services 
like RS.

I just found this even to be a bit odd.  As if they aren't really 
monitoring everything....

Chris

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ale.org/pipermail/ale/attachments/20140310/a383a7f3/attachment.html>