I agree that the convention of using -<iface> is an unfortunate choice and
prefer the concept of using a different DNS suffix insstead (it's confusing
when 'hostname' set by dhcp from different interfaces would be different,
though the hostname for a node should definitely be only one of them).
Besides, having two different IP subnets in the same forward lookup zone but
different reverse zones doesn't seem as good in this context.
In general I don't think it's good that we have any such scheme baked in, and
hope to shore that up in the future.
From: Samveen ListReceipient [mailto:[email protected]]
Sent: Thursday, July 02, 2015 7:59 AM
To: xCAT Users Mailing list
Subject: Re: [xcat-user] Static assignment of IPs
(Resending with corrections)
After more experimentation, I found a working solution for my problem.
dhcpd.conf additions:
update-static-leases on;
dhcpd.leases changes: Change the value of the option supersede
server.ddns-hostname = "xcatsn01-m"; for each host entry to match the name of
the host entry itself:
host compute1-m {
dynamic;
hardware ethernet 00:25:90:14:f3:49;
fixed-address 10.1.240.50;
supersede server.ddns-hostname = "compute1-m";
supersede host-name = "compute1";
....
}
The dhcpd.leases change is minor and shouldn't affect current installations
much (if at all).
However, the change in the dhcpd.conf is a big change in the DHCP server's
behaviour, especially for static leases. Conversely this brings up the question
that, as updates for static leases are disabled, why is there a need for the
`ddns-hostname` option in static leases?
I would love to see my use case supported by xCAT, but I am unsure of the wider
implications of the change to existing installations (and possible upgrade
paths). (I am working on a local set of changes which I want to contribute back
upstream, but depends upon the implications)
On Thu, Jul 2, 2015 at 6:55 PM, Samveen ListReceipient
<[email protected]<mailto:[email protected]>> wrote:
<snipped>
On Thu, Jul 2, 2015 at 4:07 PM, Samveen ListReceipient
<[email protected]<mailto:[email protected]>> wrote:
As per https://www.ietf.org/rfc/rfc1034.txt , the hyphen is the only
non-alphanumeric character allowed in DNS names(`.` is a separator). In fact,
the default behaviour of xCAT is to use hyphenated names for all secondary
interfaces (see the /etc/hosts output in my previous mails).
On Thu, Jul 2, 2015 at 3:31 PM, Samveen ListReceipient
<[email protected]<mailto:[email protected]>> wrote:
My problem is not with the `host-name` field of the lease (which already
matches the nodename anyway). I want the DDNS-hostname to be set to match the
node definition in xcat (and what would be generated by makedns). I have
trouble because the DHCPD incorrectly updates the hostname's A record in DDNS
to point to the last issued lease instead of the lease for the primary
interface (which DHCPD doesn't know about, and xCAT does).
If for example a BMC reset happens, then the DNS A record for the host now
points to the IPMI interface's IP.
A good example is that the hostname resolution between /etc/hosts and the DNS
server give different results:
[root@xcatmn01 xcat_prep]# host xcatsn01 localhost
Using domain server:
Name: localhost
Address: 127.0.0.1#53
Aliases:
xcatsn01.poc.lzd.co<http://xcatsn01.poc.lzd.co> has address 10.2.72.11
[root@xcatmn01 xcat_prep]# grep xcatsn01 /etc/hosts
10.4.0.10 xcatsn01 xcatsn01.poc.lzd.co<http://xcatsn01.poc.lzd.co> xcatsn01-eth1
10.1.240.10 xcatsn01-m xcatsn01-m.poc.lzd.co<http://xcatsn01-m.poc.lzd.co>
10.2.72.11 xcatsn01-s xcatsn01-s.poc.lzd.co<http://xcatsn01-s.poc.lzd.co>
In my setup, I depend on DHCPD to set the correct DNS names instead of using
`makedns xcatsn01`, and current behaviour doesn't match expected behaviour.
On Thu, Jul 2, 2015 at 12:52 PM, Xiao Peng Wang
<[email protected]<mailto:[email protected]>> wrote:
It will cause problem that your node name is different with the
hostname/ddns-name in the dhcp leases file. Are both the hostname xcatsn01-s
and xcatsn01 exist? You should try to figure out the possible error in the dns
resolution. And avoid using '-' in node name is a good idea.
Thanks
Best Regards
----------------------------------------------------------------------
Wang Xiaopeng (王晓朋)
IBM China System Technology Laboratory
Tel: 86-10-82453455
Email: [email protected]<mailto:[email protected]>
Address: 28,ZhongGuanCun Software Park,No.8 Dong Bei Wang West Road, Haidian
District Beijing P.R.China 100193
[Inactive hide details for Samveen ListReceipient ---2015/07/01 13:04:43---I
forgot to add that for the interface bound to xcats]Samveen ListReceipient
---2015/07/01 13:04:43---I forgot to add that for the interface bound to
xcatsn01-s, the complete network is the dynamic rang
From: Samveen ListReceipient <[email protected]<mailto:[email protected]>>
To: xCAT
<[email protected]<mailto:[email protected]>>
Date: 2015/07/01 13:04
Subject: Re: [xcat-user] Static assignment of IPs
________________________________
I forgot to add that for the interface bound to xcatsn01-s, the complete
network is the dynamic range:
[root@xcatmn01 xcat_prep]# lsdef -t network PROVISIONING-SVC -i dynamicrange
Object name: PROVISIONING-SVC
dynamicrange=10.2.72.2-10.2.72.254
On Wed, Jul 1, 2015 at 12:01 PM, Samveen ListReceipient
<[email protected]<mailto:[email protected]>> wrote:
The network related items of my node definition are as follows:
[root@xcatmn01 xcat_prep]# lsdef -t node -o xcatsn01 -i
installnic,ip,mac,nicips,nictypes,primarynic,nichostnamesuffixes
Object name: xcatsn01
installnic=eth0
ip=10.4.0.10
mac=00:25:90:14:f8:18!xcatsn01-s|00:25:90:14:f8:19!xcatsn01|00:25:90:14:f2:c3!xcatsn01-m
nichostnamesuffixes.ipmi=-m
nichostnamesuffixes.eth0=-s
nicips.eth1=10.4.0.10
nicips.ipmi=10.1.240.10
nicips.eth0=10.2.72.4
nictypes.ipmi=Ethernet
nictypes.eth1=Ethernet
nictypes.eth0=Ethernet
primarynic=eth1
The /etc/hosts entries are as follows:
[root@xcatmn01 xcat_prep]# grep xcatsn01 /etc/hosts
10.4.0.10 xcatsn01 xcatsn01.poc.lzd.co<http://xcatsn01.poc.lzd.co/>
xcatsn01-eth1
10.1.240.10 xcatsn01-m xcatsn01-m.poc.lzd.co<http://xcatsn01-m.poc.lzd.co/>
10.2.72.4 xcatsn01-s xcatsn01-s.poc.lzd.co<http://xcatsn01-s.poc.lzd.co/>
The DHCP lease entries are as follows:
host xcatsn01 {
dynamic;
hardware ethernet 00:25:90:14:f8:19;
fixed-address 10.4.0.10;
supersede server.ddns-hostname = "xcatsn01";
supersede host-name = "xcatsn01";
if option user-class-identifier = "xNBA" and option
client-architecture = 00:00 {
supersede server.filename =
"http://${next-server}/tftpboot/xcat/xnba/nodes/xcatsn01<http://$%7bnext-server%7d/tftpboot/xcat/xnba/nodes/xcatsn01>";
} elsif option client-architecture = 00:00 {
supersede server.filename = "xcat/xnba.kpxe";
} else {
supersede server.filename = "";
}
}
host xcatsn01-m {
dynamic;
hardware ethernet 00:25:90:14:f2:c3;
fixed-address 10.1.240.10;
supersede server.ddns-hostname = "xcatsn01";
supersede host-name = "xcatsn01";
if option user-class-identifier = "xNBA" and option
client-architecture = 00:00 {
supersede server.filename =
"http://${next-server}/tftpboot/xcat/xnba/nodes/xcatsn01<http://$%7bnext-server%7d/tftpboot/xcat/xnba/nodes/xcatsn01>";
} elsif option client-architecture = 00:00 {
supersede server.filename = "xcat/xnba.kpxe";
} else {
supersede server.filename = "";
}
}
host xcatsn01-s {
dynamic;
hardware ethernet 00:25:90:14:f8:18;
supersede server.ddns-hostname = "xcatsn01";
supersede host-name = "xcatsn01";
if option user-class-identifier = "xNBA" and option
client-architecture = 00:00 {
supersede server.filename =
"http://${next-server}/tftpboot/xcat/xnba/nodes/xcatsn01<http://$%7bnext-server%7d/tftpboot/xcat/xnba/nodes/xcatsn01>";
} elsif option client-architecture = 00:00 {
supersede server.filename = "xcat/xnba.kpxe";
} else {
supersede server.filename = "";
}
}
As can be seen from the lease entry for xcatsn01-s and xcatsn01-m, the
ddns-hostname is the same for all 3 entries. My expectation of the ddns
behaviour is that it should match the /etc/hosts (and `makedns`) actions. ie.
for xcatsn01-s, the server.ddns-hostname = "xcatsn01-s"; and for xcatsn01-m,
the server.ddns-hostname = "xcatsn01-m";
The existing behaviour causes trouble for me as the IP for xcatsn01-s is
assigned from the dynamic range, instead of the static range, and thus my
/etc/hosts and DNS entries don't remain in sync, leading to unexpected
behaviour.
Am I correct in my expectation or it is unrealistic?
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of ..)
--
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of
..)------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/_______________________________________________
xCAT-user mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/xcat-user
------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
xCAT-user mailing list
[email protected]<mailto:[email protected]>
https://lists.sourceforge.net/lists/listinfo/xcat-user
--
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of ..)
--
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of ..)
--
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of ..)
--
--
Samveen S. Gulati
http://samveen.in<http://samveen.in/>
The best-laid schemes o' mice an 'men
Gang aft agley,
An'lea'e us nought but grief an' pain,
For promis'd joy!
-- Robert Burns
(The best laid plans of mice and men often go awry,
and bring nothing but grief and pain instead of ..)
------------------------------------------------------------------------------
Don't Limit Your Business. Reach for the Cloud.
GigeNET's Cloud Solutions provide you with the tools and support that
you need to offload your IT needs and focus on growing your business.
Configured For All Businesses. Start Your Cloud Today.
https://www.gigenetcloud.com/
_______________________________________________
xCAT-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xcat-user
