cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Caleb Call <calebc...@me.com>
Subject Re: alter resolv.conf nameservers on linux
Date Thu, 08 Nov 2012 18:46:22 GMT
Yes, I restarted it through the UI. (Actually, I stopped it then started it).

/var/cache/cloud/cmdline last option set is useextdns=true (if you want more from this output,
let me know)

root@r-4-VM:~# grep "dhcp-option=6" /etc/dnsmasq.conf
dhcp-option=6,<dns1 ip addr>,<dn2 ip addr>

[root@management cloud-management]# zgrep "Boot Args for" /var/log/cloud/management/management-server.log.2012-11-07.gz

2012-11-07 16:50:03,188 DEBUG [network.router.VirtualNetworkApplianceManagerImpl] (Job-Executor-78:job-505)
Boot Args for VM[DomainRouter|r-4-VM]:  template=domP type=dhcpsrvr  name=r-4-VM eth0ip=<ip
addr> eth0mask=255.255.255.0 gateway=<ip addr> eth1ip=169.254.1.60 eth1mask=255.255.0.0
dhcprange=<ip addr> domain=cs1cloud.internal dns1=<dns1 ip addr> dns2=<dns2
ip addr> useextdns=true
2012-11-07 22:28:04,483 DEBUG [network.router.VirtualNetworkApplianceManagerImpl] (Job-Executor-1:job-506)
Boot Args for VM[DomainRouter|r-4-VM]:  template=domP name=r-4-VM eth0ip=<ip addr> eth0mask=255.255.255.0
gateway=<ip addr> domain=cs1cloud.internal dhcprange=<ip addr> eth1ip=169.254.3.22
eth1mask=255.255.0.0 type=dhcpsrvr disable_rp_filter=true dns1=<dns1 ip addr> dns2=<dns2
ip addr> useextdns=true
2012-11-07 22:35:52,482 DEBUG [storage.secondary.SecondaryStorageManagerImpl] (Job-Executor-3:job-508)
Boot Args for VM[SecondaryStorageVm|s-256-cv]:  template=domP type=secstorage host=<ip
addr> port=8250 name=s-256-cv zone=1 pod=1 guid=s-256-cv resource=com.cloud.storage.resource.NfsSecondaryStorageResource
instance=SecStorage sslcopy=true role=templateProcessor mtu=1500 eth2ip=209.41.84.81 eth2mask=255.255.255.0
gateway=<ip addr> eth0ip=169.254.1.29 eth0mask=255.255.0.0 eth1ip=<ip addr> eth1mask=255.255.255.0
mgmtcidr=<ip addr>/24 localgw=<ip addr> private.network.device=eth1 eth3ip=<ip
addr> eth3mask=255.255.255.0 storageip=<ip addr> storagenetmask=255.255.255.0 storagegateway=<ip
addr> internaldns1=<dns1 ip addr> internaldns2=<dns2 ip addr> dns1=<dns1
ip addr> dns2=<dns2 ip addr>
2012-11-07 22:37:34,906 DEBUG [cloud.consoleproxy.ConsoleProxyManagerImpl] (Job-Executor-5:job-510)
Boot Args for VM[ConsoleProxy|v-2-VM]:  template=domP type=consoleproxy host=<ip addr>
port=8250 name=v-2-VM premium=true zone=1 pod=1 guid=Proxy.2 proxy_vm=2 disable_rp_filter=true
eth2ip=<ip addr>23 eth2mask=255.255.255.0 gateway=<ip addr> eth0ip=169.254.3.72
eth0mask=255.255.0.0 eth1ip=<ip addr> eth1mask=255.255.255.0 mgmtcidr=<ip addr>/24
localgw=<ip addr> internaldns1=<dns1 ip addr> internaldns2=<dns2 ip addr>
dns1=<dns1 ip addr> dns2=<dns2 ip addr>
2012-11-07 22:39:12,634 DEBUG [storage.secondary.SecondaryStorageManagerImpl] (Job-Executor-7:job-512)
Boot Args for VM[SecondaryStorageVm|s-257-VM]:  template=domP type=secstorage host=<ip
addr> port=8250 name=s-257-VM zone=1 pod=1 guid=s-257-VM resource=com.cloud.storage.resource.NfsSecondaryStorageResource
instance=SecStorage sslcopy=true role=commandExecutor mtu=1500 eth2ip=<ip addr> eth2mask=255.255.255.0
gateway=<ip addr> eth0ip=169.254.1.112 eth0mask=255.255.0.0 eth1ip=<ip addr> eth1mask=255.255.255.0
mgmtcidr=<ip addr>/24 localgw=<ip addr> private.network.device=eth1 eth3ip=<ip
addr> eth3mask=255.255.255.0 storageip=<ip addr> storagenetmask=255.255.255.0 storagegateway=<ip
addr> internaldns1=<dns1 ip addr> internaldns2=<dns2 ip addr> dns1=<dns1
ip addr> dns2=<dns2 ip addr>
[root@management cloud-management]#


On Nov 8, 2012, at 11:18 AM, Chiradeep Vittal <Chiradeep.Vittal@citrix.com> wrote:

> Did you restart the VR through the api / UI ?
> What is the output of
> cat /var/cache/cloud/cmdline
> grep "dhcp-option=6" /etc/dnsmasq.conf
> 
> Management server log:
> grep "Boot Args for" <log file>
> 
> On 11/7/12 8:40 PM, "Caleb Call" <calebcall@me.com> wrote:
> 
>> Sorry, yes, setting the global parameter does not work.  The contents of
>> my /etc/resolv.conf is:
>> 
>> # cat /etc/resolv.conf
>> ; generated by /sbin/dhclient-script
>> search cs1cloud.internal
>> nameserver <ip of my vrouter>
>> 
>> I can edit it to what I want it to be, restart the instance and it gets
>> changed back to this.  This happens across all my zones in all my
>> different Cloudstack environments.  DNS is configured correctly for them.
>> 
>> The reason for my desire to not use my virtual router is two fold.
>> First, I have found the name resolution (forwarding if you will) through
>> the VR to be spotty at best.  To correct it, I usually have to bounce my
>> VR, it's like it just stops passing traffic.  This is happening in both
>> of my Cloudstack environments.  Second, running a single nameserver is by
>> no means acceptable in any environment.  I wouldn't do it in my dev,
>> test, or stage environments let alone my production environment.  We also
>> specify additional parameters in /etc/resolv.conf like timeout and we
>> have to be able to specify additional search domains.
>> 
>> 
>> On Nov 7, 2012, at 6:07 PM, Chiradeep Vittal
>> <Chiradeep.Vittal@citrix.com> wrote:
>> 
>>> Not sure what you mean by #2. Do you mean that you set the global
>>> configuration flag use.external.dns ?
>>> If so, what is the content of /etc/resolv.conf ?
>>> 
>>> Why is the DNS forwarder not working for you? Is your zone dns
>>> configured
>>> correctly in CloudStack?
>>> 
>>> On 11/7/12 4:06 PM, "Caleb Call" <calebcall@me.com> wrote:
>>> 
>>>> #2 doesn't work.  It still stomps /etc/resolv.conf with what's provided
>>>> by the routervm.
>>>> 
>>>> On Nov 7, 2012, at 4:48 PM, Caleb Call <calebcall@me.com> wrote:
>>>> 
>>>>> Thanks, #2 looks perfect.  I'll give that a shot.
>>>>> 
>>>>> On Nov 7, 2012, at 4:29 PM, Chiradeep Vittal
>>>>> <Chiradeep.Vittal@citrix.com> wrote:
>>>>> 
>>>>>> That should not be necessary. The DNS server in the router is a
>>>>>> forwarder:
>>>>>> 1. If the target of the DNS resolve is for a VM it has served DHCP
>>>>>> to,
>>>>>> it
>>>>>> responds with the entry
>>>>>> 2. If not, it forwards it to the 'zone' dns server.
>>>>>> 
>>>>>> Or, you can set use.external.dns to 'true', restart the management
>>>>>> server
>>>>>> and restart (from the api/ui) the virtual router.
>>>>>> But if you do, you won't get #1.
>>>>>> 
>>>>>> On 11/7/12 3:12 PM, "Alex Huang" <Alex.Huang@citrix.com> wrote:
>>>>>> 
>>>>>>> In 4.0, what I would do is this.
>>>>>>> 
>>>>>>> - Write a plugin.
>>>>>>> - Listen to vm start events.
>>>>>>> - On router vm start, ssh into the router vm and change the
>>>>>>> resolv.conf
>>>>>>> 
>>>>>>> In the next release, Murali have added proper external event
system
>>>>>>> then
>>>>>>> you don't even need to do this via a plugin.
>>>>>>> 
>>>>>>> --Alex
>>>>>>> 
>>>>>>> -----Original Message-----
>>>>>>> From: Caleb Call [mailto:calebcall@me.com]
>>>>>>> Sent: Wednesday, November 07, 2012 3:06 PM
>>>>>>> To: cloudstack-users@incubator.apache.org
>>>>>>> Subject: Re: alter resolv.conf nameservers on linux
>>>>>>> 
>>>>>>> Good question, I've been meaning to ask this same thing and keep
>>>>>>> forgetting to.  I think I've read that you have to edit the config
>>>>>>> on
>>>>>>> the
>>>>>>> router vm, but that doesn't persist a reboot of the router vm.
 Is
>>>>>>> there
>>>>>>> a better way to do this?
>>>>>>> 
>>>>>>> 
>>>>>>> On Nov 7, 2012, at 3:18 PM, "Musayev, Ilya" <imusayev@webmd.net>
>>>>>>> wrote:
>>>>>>> 
>>>>>>>> How would I pass on my nameservers in resolv.conf of an instance,
>>>>>>>> instead of router vms IP?
>>>>>>>> 
>>>>>>>> As of now, my nameserver is set to ip address of router vm.
>>>>>>>> 
>>>>>>>> 
>>>>>>>> Thanks
>>>>>>>> ilya
>>>>>>> 
>>>>>> 
>>>>> 
>>>> 
>>> 
>> 
> 


Mime
View raw message