You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@cloudstack.apache.org by "Nuno Tavares (JIRA)" <ji...@apache.org> on 2015/01/21 15:18:35 UTC

[jira] [Comment Edited] (CLOUDSTACK-7028) [RVR] Static NAT does not work after the fail-over in additional public range

    [ https://issues.apache.org/jira/browse/CLOUDSTACK-7028?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14285633#comment-14285633 ] 

Nuno Tavares edited comment on CLOUDSTACK-7028 at 1/21/15 2:17 PM:
-------------------------------------------------------------------

Hi Jayapal,

Although there are not "practical" issues, as the packets routed through eth2 will eventually default to the 'main' default gateway, I believe that, for consistency' sake, the script should handle Table_eth2 as well for this particular issue, since you're doing it everywhere. You are:
- marking (mangling) the packets for eth2
- adding ip rules for table lookups for eth2
- defining a table for eth2 (Table_eth2)

... right?

Furthermore, ignoring eth2 specifically means that CS would be basing logic in an assumption (described above) which is not good, from my point of view; so if we think that eth2 belongs to a group of interfaces such as eth3 does, then I find not reason to treat it differently.



was (Author: ntavares):
Hi Jaypal,

Although there are not "practical" issues, as the packets routed through eth2 will eventually default to the 'main' default gateway, I believe that, for consistency' sake, the script should handle Table_eth2 as well for this particular issue, since you're doing it everywhere. You are:
- marking (mangling) the packets for eth2
- adding ip rules for table lookups for eth2
- defining a table for eth2 (Table_eth2)

... right?

Furthermore, ignoring eth2 specifically means that CS would be basing logic in an assumption (described above) which is not good, from my point of view; so if we think that eth2 belongs to a group of interfaces such as eth3 does, then I find not reason to treat it differently.


> [RVR] Static NAT does not work after the fail-over in additional public range
> -----------------------------------------------------------------------------
>
>                 Key: CLOUDSTACK-7028
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7028
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the default.) 
>          Components: Network Controller
>    Affects Versions: 4.0.0
>            Reporter: Jayapal Reddy
>            Assignee: Jayapal Reddy
>              Labels: AUTOMATION_REQ, DEVREV
>             Fix For: 4.4.0
>
>
> On fail over, in master router route got missed.
> Reproducing steps:
> 1. Create RVR network and acquire additional public ip range (ex: 47 vlan and 10.147.47.x subnet)
> 2. create a static nat rue on additional range public ip and add firewall rule for port 22-22
> 3. ssh to public ip, it get connected to vm
> 4. Now make master VR down, backup wil become master.
> 5. On master router on eth3 there default route got missed, which is causing the ingress traffic is coming to eth3 is going out via eth2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)