Re: Private Gateway on REDUNDANT VPC

[Simon Weller]

Paul,


Could it be related to rp_filter in some way?


- Si



From: Paul Angus 
Sent: Tuesday, June 20, 2017 3:39 AM
To: dev@cloudstack.apache.org
Subject: RE: Private Gateway on REDUNDANT VPC

I don't believe so.
Rules look OK and consistent with std VPC as well.


Kind regards,

Paul Angus

paul.an...@shapeblue.com
www.shapeblue.com
[http://shapeblue.com/wp-content/uploads/2014/03/sungardonline1.jpg]

Shapeblue - The CloudStack Company
www.shapeblue.com
The city of Prague was the venue for the spring meeting of the Cloudstack 
European user group. There was



53 Chandos Place, Covent Garden, London  WC2N 4HSUK
@shapeblue




-Original Message-
From: Jayapal Uradi [mailto:jayapal.ur...@accelerite.com]
Sent: 20 June 2017 09:30
To: dev@cloudstack.apache.org
Subject: Re: Private Gateway on REDUNDANT VPC

Did you check iptables, Is it blocking on the VR ?

> On Jun 20, 2017, at 1:30 PM, Paul Angus  wrote:
>
> Hi All,
>
> I've been looking at the failing Marvin tests for Private Gateways.It 
> passes on std VPC and fails on rVPC.
> The test tries to ping a VM on a remote VPC via the private gateways on both 
> VRs.
> Digging into it, I found that an ARP request goes out for the remote VM from 
> the local VR to the remote VR, the local VR receives it, then nothing.  On 
> the std VRs a reply goes back out.
>
> I've checked all interfaces to see if the reply is going out of the wrong 
> interface, but it just isn't going out anywhere.  I can't figure out why no 
> reply seems to be generated...  Obviously the answer is in the difference in 
> config and packages on VPC vs rVPC - but I can't find it.
>
> HELP!  Any ideas anyone?
>
>
> Kind regards,
>
> Paul Angus
>
>
> paul.an...@shapeblue.com
> www.shapeblue.com
> 53 Chandos Place, Covent Garden, London  WC2N 4HSUK @shapeblue
>
>
>

DISCLAIMER
==
This e-mail may contain privileged and confidential information which is the 
property of Accelerite, a Persistent Systems business. It is intended only for 
the use of the individual or entity to which it is addressed. If you are not 
the intended recipient, you are not authorized to read, retain, copy, print, 
distribute or use this message. If you have received this communication in 
error, please notify the sender and delete all copies of this message. 
Accelerite, a Persistent Systems business does not accept any liability for 
virus infected mails.

Re: [VOTE] Apache CloudStack 4.10.0.0 RC5

[Rajani Karuturi]

Agree. Will create rc6 soon.

~Rajani

Sent from phone.

On 23 Jun 2017 3:17 p.m., "Wido den Hollander"  wrote:

>
> > Op 22 juni 2017 om 21:38 schreef Kris Sterckx <
> kris.ster...@nuagenetworks.net>:
> >
> >
> > Sorry , i meant  PR/2084  >
> >
>
> Ok, so revert and RC6 then?
>
> Wido
>
> > On 22 June 2017 at 21:25, Kris Sterckx 
> > wrote:
> >
> > > Hi all
> > >
> > >
> > > Thanks to Daan for driving the work at PR/2155
> > >  , but i believe it
> needs
> > > more soak time.
> > >
> > > We found today this PR is breaking VR-provided Guest DNS for SDN
> providers
> > > also.
> > >
> > > As Daan also suggested, i would recommend taking it out 4.10 and give
> it
> > > some more cycles in 4.11.
> > >
> > > We will support qualifying it.
> > >
> > > thanks,
> > >
> > > Kris
> > >
> > > - Nuage Networks
> > >
> > > On 21 June 2017 at 16:45, Kris Sterckx  >
> > > wrote:
> > >
> > >> Hi  Wido, Daan, all
> > >>
> > >>
> > >> We have just pushed a workaround
> > >>
> > >> https://github.com/apache/cloudstack/pull/2155
> > >>
> > >> This works for us.
> > >>
> > >> The CsPassword class seems like not needed at all ?  The important
> part
> > >> for us is that  iptables_change = True  is set also in case of
> password
> > >> handling.  That was missing.
> > >>
> > >> Daan, we left a TODO for you to look into deeper though. We worked at
> > >> getting the databag handling fixed inside the class but as we don't
> have
> > >> full background, we kept hitting issues. Until we then saw that
> commenting
> > >> out the lines which we did actually works as well.
> > >>
> > >> I suggest the /2155 is reviewed by the involved stakeholders and
> based on
> > >> that we decide ?
> > >>
> > >>
> > >> thanks and toi-toi :-)
> > >>
> > >>
> > >> Kris
> > >>
> > >>
> > >>
> > >>
> > >
> > >
>

Re: [VOTE] Apache CloudStack 4.10.0.0 RC5

[Wido den Hollander]

> Op 22 juni 2017 om 21:38 schreef Kris Sterckx 
> :
> 
> 
> Sorry , i meant  PR/2084 
> 

Ok, so revert and RC6 then?

Wido

> On 22 June 2017 at 21:25, Kris Sterckx 
> wrote:
> 
> > Hi all
> >
> >
> > Thanks to Daan for driving the work at PR/2155
> >  , but i believe it needs
> > more soak time.
> >
> > We found today this PR is breaking VR-provided Guest DNS for SDN providers
> > also.
> >
> > As Daan also suggested, i would recommend taking it out 4.10 and give it
> > some more cycles in 4.11.
> >
> > We will support qualifying it.
> >
> > thanks,
> >
> > Kris
> >
> > - Nuage Networks
> >
> > On 21 June 2017 at 16:45, Kris Sterckx 
> > wrote:
> >
> >> Hi  Wido, Daan, all
> >>
> >>
> >> We have just pushed a workaround
> >>
> >> https://github.com/apache/cloudstack/pull/2155
> >>
> >> This works for us.
> >>
> >> The CsPassword class seems like not needed at all ?  The important part
> >> for us is that  iptables_change = True  is set also in case of password
> >> handling.  That was missing.
> >>
> >> Daan, we left a TODO for you to look into deeper though. We worked at
> >> getting the databag handling fixed inside the class but as we don't have
> >> full background, we kept hitting issues. Until we then saw that commenting
> >> out the lines which we did actually works as well.
> >>
> >> I suggest the /2155 is reviewed by the involved stakeholders and based on
> >> that we decide ?
> >>
> >>
> >> thanks and toi-toi :-)
> >>
> >>
> >> Kris
> >>
> >>
> >>
> >>
> >
> >