1. Cf-Dev
  2. Topics

xip.io IO errors

Sumanth Yamala
 

Hi,

Currently we see IO errors when trying to contact xip.io - this happens around 2% of the times. To apply a temporary fix - I updated the /etc/hosts of the ha_proxy VM to map the - IP to the service routes of each app deployed on the system.

After the above change I still see the packet drops - so wanted to know if this needs to be done on all the runners as well or ..? any thoughts.
Thanks,
Sumanth

attachment.html

attachment.html

Amit Kumar Gupta
 

Yes, xip.io is flaky. I would recommend setting up real DNS if you want to
avoid the 2% failure rate if you want a longer term, reliable solution.

On Thu, Oct 29, 2015 at 4:55 PM, Sumanth Yamala <Sumanth.Yamala(a)sas.com>
wrote:

Hi,



Currently we see IO errors when trying to contact xip.io – this happens
around 2% of the times. To apply a temporary fix – I updated the /etc/hosts
of the ha_proxy VM to map the - IP to the service routes of each app
deployed on the system.



After the above change I still see the packet drops – so wanted to know if
this needs to be done on all the runners as well or ..? any thoughts.

Thanks,

Sumanth

attachment.html

attachment.html

Dan Wendorf
 

If you're using the common bosh lite IP of 10.244.0.34, you can also use
the more-reliable *.bosh-lite.com

On Thu, Oct 29, 2015, 5:40 PM Amit Gupta <agupta(a)pivotal.io> wrote:

Yes, xip.io is flaky. I would recommend setting up real DNS if you want
to avoid the 2% failure rate if you want a longer term, reliable solution.

On Thu, Oct 29, 2015 at 4:55 PM, Sumanth Yamala <Sumanth.Yamala(a)sas.com>
wrote:

Hi,



Currently we see IO errors when trying to contact xip.io – this happens
around 2% of the times. To apply a temporary fix – I updated the /etc/hosts
of the ha_proxy VM to map the - IP to the service routes of each app
deployed on the system.



After the above change I still see the packet drops – so wanted to know
if this needs to be done on all the runners as well or ..? any thoughts.

Thanks,

Sumanth

attachment.html

attachment.html

Juan Antonio Breña Moral <bren at juanantonio.info...>
 

I was reading this thread, I noticed that I have the same problem and I thought that we had a problem in our network.

In local I am using development from Yudai:
https://github.com/yudai/cf_nise_installer

I am going to fix this issue too
Guillaume Berche
 

Would the community release embedding power dns to provide xip.io style
impl be appliceable to your contexte and useful [1] ? This relies a DNS
delegation though. I haven't tested it myself yet.

[1] https://github.com/cloudfoundry-community/xip-release

Guillaume.

Le 30 oct. 2015 01:45, "Dan Wendorf" <dwendorf(a)pivotal.io> a écrit :

If you're using the common bosh lite IP of 10.244.0.34, you can also use
the more-reliable *.bosh-lite.com

On Thu, Oct 29, 2015, 5:40 PM Amit Gupta <agupta(a)pivotal.io> wrote:

Yes, xip.io is flaky. I would recommend setting up real DNS if you want
to avoid the 2% failure rate if you want a longer term, reliable solution.

On Thu, Oct 29, 2015 at 4:55 PM, Sumanth Yamala <Sumanth.Yamala(a)sas.com>
wrote:

Hi,



Currently we see IO errors when trying to contact xip.io – this happens
around 2% of the times. To apply a temporary fix – I updated the /etc/hosts
of the ha_proxy VM to map the - IP to the service routes of each app
deployed on the system.



After the above change I still see the packet drops – so wanted to know
if this needs to be done on all the runners as well or ..? any thoughts.

Thanks,

Sumanth

attachment.html

attachment.html

1 - 5 of 5