I actually hacked up a wrapper to the /usr/bin/omshell that would write
data to multiple DHCP servers.
It worked and past my initial testing but I never got much of a chance
to implement it.
The requirement on my end for failover got eclipsed by other more
pressing problems.
I did post the script to [email protected] and I believe that he
was able to use it.
I could post it to the mail list in general if there is any interest.
On 02/17/2017 07:51 PM, Isabell Cowan wrote:
Not a solution per say, but a reasonable work around. Call the
attached script in a frequent cron job.
On Monday, August 15, 2016 at 11:02:50 AM UTC-5, Alvin Starr wrote:
I have looked at this a bit more and your right.
Its hard to believe that the ISC develpers would be that short
sighted to not extend the design so that changes made to the
primary are copied to the secondary.
I can see 3 possible solutions.
1) fix the ISC DHCP server.
Fixing the DHCP server would be a major pain and there may be
some fundamental reason that its not possible to copy the static
entries
2) change dhcp_isc_main.rb to contact both primary and secondary
DHCP servers.
Looking at the code it would be possible to change the omcmd
routine to talk to a secondary server.
This would require adding configuration items to handle the
secondary.
The problem I can see here is what to do in the case of an
error condition from one server and not the other?
3) replace omshell with a shell that writes to both servers.
An omshell wrapper could parse the dhcpd.conf file and
determine if there is a secondary server.
If there is a secondary server the commands could be piped to
both servers.
Once again there is the issue of what to do about errors in
one server and not the other.
This solution would also likely require a small patch to
dhcp_isc_main.rb to allow for changing.
I could supply a patch to the dhcp_isc_main.rb and do some testing
on my environment so that it would pass the initial sniff test but
I am not in a position to do a full pull request and then submit
the patches via usual developer channel.
On 08/11/2016 06:40 PM, [email protected] <javascript:> wrote:
DHCP will not replicate the static leases as I said before, you
need twice the omapi using omshell.
I have discussed this with the ISC guys.
Op vrijdag 12 augustus 2016 00:26:15 UTC+2 schreef Alvin Starr:
We have slightly different problems but are both limited by
the nature of foremans dhcp-proxy.
If you create static leases using omshell then I believe the
data will be replicated to a failover dhcp server.
Forman will create new hosts using omshell.
The comment about DNS is only because foreman can cleanly
interact with DNS remotely.
On 08/11/2016 06:05 PM, [email protected] wrote:
You are completely wrong.
What we like to have is a second DHCP with a proxy where is
talked to using OMAPI by foreman so we have a replication
about the static leases.
DHCP itself is unable to sync them this way and for Failover
we need to have 2 writes to 2 DHCP servers which are in
clutser mode.
DNS is not involved here.
Op donderdag 11 augustus 2016 23:50:43 UTC+2 schreef Alvin
Starr:
A standalone DHCP server without lots of extra software
installed.
I do not want to install the whole forman-proxy on the
DHCP or DNS servers
As an asidethe isc_dhcp proxy does not parse correctly
formatted dhcpd.conffiles(I just filed a bug report).
Looking at omapi I could be convinced that it is
impossible to get thecurrent dhcp config
informationremotely.
On 08/11/2016 05:26 PM, [email protected] wrote:
What do you mean by a standalone DHCP server ? The
proxy handles that when you installed the proxy on the
DHCP server.
This goes about a second DHCP server that knows the
same static leases as the proxy knows for failover
whent the primary, with proxy, fails.
Op donderdag 11 augustus 2016 22:39:40 UTC+2 schreef
Alvin Starr:
I have wondered about the same problem.
Since foreman needs to read and write the leases
and config files it means that you cannot have a
remote standalone DHCP server.
On 08/11/2016 04:32 PM, [email protected] wrote:
Is this me or does no-one care about DHCP redundancy ?
Op zaterdag 16 januari 2016 20:06:30 UTC+1 schreef
Matt .:
Hi,
I was wondering if someone has a workaround
for DHCP sync with failover.
As foreman writes to the leases file directly
this is an issue for syncing.
Can't we write to a config file we include to
the dhcpd.conf and
reload dhcp when the proxy did something to
dhcp ?
Ideas are welcome.
Cheers,
Matt
--
You received this message because you are
subscribed to the Google Groups "Foreman users" group.
To unsubscribe from this group and stop receiving
emails from it, send an email to
[email protected].
To post to this group, send email to
[email protected].
Visit this group at
https://groups.google.com/group/foreman-users.
For more options, visit
https://groups.google.com/d/optout.
--
Alvin Starr || voice: (905)513-7688
Netvel Inc. || Cell: (416)806-0133
[email protected] ||
--
Alvin Starr || voice: (905)513-7688
Netvel Inc. || Cell: (416)806-0133
[email protected] ||
--
Alvin Starr || voice: (905)513-7688
Netvel Inc. || Cell: (416)806-0133
[email protected] ||
--
Alvin Starr || voice: (905)513-7688
Netvel Inc. || Cell: (416)806-0133
[email protected] <javascript:> ||
--
Alvin Starr || voice: (905)513-7688
Netvel Inc. || Cell: (416)806-0133
[email protected] ||
--
You received this message because you are subscribed to the Google Groups "Foreman
users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/foreman-users.
For more options, visit https://groups.google.com/d/optout.
