cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From John Pletka <jple...@abraxis.com>
Subject Re: Management IP on guest VM using public IP
Date Fri, 14 Nov 2014 23:48:40 GMT
My temporary work-around is to setup a cron job to delete that bad
route when the server restarts, but I'm trying to find a primary fix.

Management server public IP:  xx.47.90.4
Management server private IP: 10.1.40.3

Currently, the management.network.cidr is set to xx.47.90.0/24.
Should that be changed to 10.1.40.0/24?

If not, any suggestions on stopping that bad route from being send to
the infrastructure VMs?

On Fri, Oct 24, 2014 at 10:31 AM, John Pletka <jpletka@abraxis.com> wrote:
> My internal DNS is working (I installed dnsmasq on the management server and
> set it to listen on the internal IP address 10.1.40.3).  From the hosts and
> guests, pinging the management server by name goes to the correct internal
> address.  Where do the arguements in /var/cache/cloud/cmdline come from?
> That appears to be the source of the problem.
> mgmtcidr=XX.47.90.0/24 localgw=10.1.40.1
> host=xx.47.90.4
>
> The localgw is correct, but the mgmtcidr and host are wrong (they point to
> public instead of local)
>
> John
>
>
>
> On Fri, Oct 24, 2014 at 10:05 AM, Andrija Panic <andrija.panic@gmail.com>
> wrote:
>>
>> and internal DNS - it must be accessible/routable from mgmt network - when
>> I set internal dns to be 8.8.8.8 then SSVM check was failing, and I had
>> bad
>> routes as you do at the moment...
>> So no routes for internal DNS should be added at all in my opinion, it's
>> is
>> on the administarator of the CS to allow public IP (whatever it is private
>> ro public IP) to be reachable/routable from mgmt network...
>>
>> On 24 October 2014 16:03, Andrija Panic <andrija.panic@gmail.com> wrote:
>>
>> > I had the same/similar issues, when my mgmt server and agents (bridge on
>> > them) had public IP...I add host as 10.x,x,x to the cloudstack, and the
>> > cs-agent reads the public IP address from the bridge and defines that
>> > public IP as the host IP inside cloudstack (acs 4.4)
>> >
>> > On 24 October 2014 01:06, John Pletka <jpletka@abraxis.com> wrote:
>> >
>> >> This is probably something simple, but I can't find it.
>> >> *TLDR: how to send guests the internal IP of the management server
>> >> instead
>> >> of public?*
>> >>
>> >> My setup:
>> >>
>> >> xx.47.90.0/24 => cloud-public
>> >> 10.1.40.0/24 => cloud-private (SAN + MGMT)
>> >>
>> >> Management server: xx.47.90.4 (public IP) and 10.1.40.4 (private IP)
>> >>
>> >> I'm trying to run the storage server health check
>> >> (/usr/local/cloud/systemvm/ssvm-check.sh) and it is failing pinging the
>> >> management server.
>> >>
>> >> Good: DNS resolves download.cloud.com
>> >> ================================================
>> >> nfs is currently mounted
>> >> ================================================
>> >> Management server is XX.47.90.4. Checking connectivity.
>> >> ERROR: Cannot connect to XX.47.90.4 port 8250
>> >> 2014/10/23 22:51:29 socat[4617] E write(3, 0x86773c8, 1): No route to
>> >> host
>> >>
>> >> *The reason it is failing is it adds a default route to the Secondary
>> >> Storage server pointing that public IP at the local gateway.*  *If I
>> >> manually delete that route, the health check works*
>> >>
>> >> root@s-6-VM:~# route -n
>> >> Kernel IP routing table
>> >> Destination     Gateway         Genmask         Flags Metric Ref    Use
>> >> Iface
>> >> 0.0.0.0         xx.47.90.1     0.0.0.0         UG    0      0        0
>> >> eth2
>> >> 10.1.40.0       0.0.0.0         255.255.255.0   U     0      0        0
>> >> eth1
>> >> 10.1.40.0       0.0.0.0         255.255.255.0   U     0      0        0
>> >> eth3
>> >> 169.254.0.0     0.0.0.0         255.255.0.0     U     0      0        0
>> >> eth0
>> >> xx.47.90.0     0.0.0.0         255.255.255.0   U     0      0        0
>> >> eth2
>> >> xx.47.90.4     10.1.40.1       255.255.255.255 UGH   0      0        0
>> >> eth1
>> >>
>> >> The test is getting that IP from /var/cache/cloud/cmdline
>> >>
>> >> root@s-6-VM:~# cat /var/cache/cloud/cmdline
>> >> root=UUID=3bbaf5c6-5317-468b-9742-0e68c65ad565 ro
>> >> debian-installer=en_US
>> >> quiet -- quiet console=hvc0 template=domP type=secstorage
>> >> host=XX.47.90.4
>> >> port=8250 name=s-6-VM zone=1 pod=1 guid=s-6-VM
>> >> resource=com.cloud.storage.resource.PremiumSecondaryStorageResource
>> >> instance=SecStorage sslcopy=false role=templateProcessor mtu=1500
>> >> eth2ip=XX.47.90.195 eth2mask=255.255.255.0 gateway=XX.47.90.1
>> >> eth0ip=169.254.2.79 eth0mask=255.255.0.0 eth1ip=10.1.40.207
>> >> eth1mask=255.255.255.0 mgmtcidr=XX.47.90.0/24 localgw=10.1.40.1
>> >> private.network.device=eth1 eth3ip=10.1.40.65 eth3mask=255.255.255.0
>> >> storageip=10.1.40.65 storagenetmask=255.255.255.0
>> >> storagegateway=10.1.40.1
>> >> internaldns1=10.1.40.3 internaldns2= dns1=XX.47.64.201
>> >> dns2=XX.47.67.201
>> >>
>> >
>> >
>> >
>> > --
>> >
>> > Andrija Panić
>> > --------------------------------------
>> >   http://admintweets.com
>> > --------------------------------------
>> >
>>
>>
>>
>> --
>>
>> Andrija Panić
>> --------------------------------------
>>   http://admintweets.com
>> --------------------------------------
>
>

Mime
View raw message