eth0: not responding (recovering)

Thu Oct 10 17:23:07 UTC 2019

Hi Bill
I non-failover DHCP server doesn't have any "failover peer" stanza in the
/var/log/dhcpd/dhcpd.leases file ---->
Sorry for typo in my previous email
The environment became most problematic none of the servers are granting
leases to devices On the primary lease file it says
my state partner-down
peer state recovery
----> Its not peer it is  partner state

 A working failover lease file has at the top ----->
I can see at multiple places in the dhcpd.lease file specifying about these
states appended with time stamp saying like
failover peer "dhcp-peer-workspace1" state {
  my state recover at *3 2019/10/09 14:23:41;*
  partner state unknown-state *at 3 2019/10/09 14:23:41;*
}

failover peer "dhcp-peer- workspace1 " state {
  my state recover at *3 2019/10/09 14:23:41;*
  partner state unknown-state at *3 2019/10/09 14:23:41*;
}
server-duid "\000\001\000\001%0\251\355\000PV\207D\342";
failover peer "dhcp-peer- workspace1 " state {
  my state recover at 3 2019/10/09 14:23:41;
  partner state unknown-state at 3 2019/10/09 14:23:41;
}
 Try shutting down both the primary and secondary servers, remove the
"failover peer" stanza---->
Yes I tried this,I shut down the failover and on primary,  I removed the
failover config part totally from the config file
Stopped the dhcpd and deleted the lease file and again touch the lease file
then restarted the DHCP it worked as expected,

The moment I bring the failover appliances up and add the failover section
to the primary config file and restart the dhcpd  on the primary  the issue
starts.
First the failover logs says it is in recovery mode ok, So i thought as it
has to sync the primary it is in recovery mode. After some span of
time on the primary appliance lease file I see
my state partner-down
partner state recovery
Thus comes the issues, and these are running in for ever condition the
status are not getting updated in any of the appliance lease file
 Is it ok if I edit the lease file manually and make it normal ?

On Thu, Oct 10, 2019 at 9:35 PM Bill Shirley <
bill at c3po.polymerindustries.biz> wrote:

> I non-failover DHCP server doesn't have any "failover peer" stanza in the
> /var/log/dhcpd/dhcpd.leases
> file.  A working failover lease file has at the top:
> failover peer "dhcp-failover" state {
>   my state normal at 4 2019/10/10 05:05:04;
>   partner state normal at 5 2011/09/02 23:51:25;
> }
>
> Try shutting down both the primary and secondary servers, remove the
> "failover peer" stanza from
> both of their lease files, and then bring up the primary *with the
> failover configuration*.  Ensure it is
> handing out leases, then bring up the secondary *with the failover
> configuration*.  Then check that
> all is working correctly.
>
> Bill
> On 10/9/2019 1:32 PM, Surya Teja wrote:
>
> I am facing weird situation with fail over setup on my lab environment. I
> am facing issue when the failover dhcp appliance is added to my existing
> server.
> For the first time when i add failover to primary appliance,  On primary
> appliance lease file i see the partner state as unknown and in the failover
> the messages are printing not responding recovery, so i shutdown the
> failover appliance and removed the failover config section from primary and
> restarted primary then it was working fine.
> As a trial of second attempt i increased mclt value to 3600 and added the
> failover section back to primary config and bring up the failover server
> now.
> The environment became most problematic none of the servers are granting
> leases to devices On the primary lease file it says
> my state partner-down
> peer state recovery
>
> Why do we get these recovery,partner down, unknown status when i add the
> failover to my environment?
> Or do we have any best practice steps how to add failover to existing
> server without causing any outages?
>
> Any help would be appreciated
> Thanks
>
>
>
> On Sun, 6 Oct 2019, 21:04 Surya Teja, <suryateja042 at gmail.com> wrote:
>
>> Hi Bill Thanks for your reply,
>> Yes I see traffic on the peer ports which i mentioned in the fail over
>> section of my configuration file.
>> My mclt value is 1800(30 min).
>> I am seeing these issues on the failover server and some times I see the
>> logs saying peer hold all free leases, but that scope is not completely
>> full with active entries in the dhcpd.lease file of that specified server
>>
>> And one more strange thing I observerd in the lease file. In the file I
>> have statements like my status and peer status. In that peer status is
>> saying *unknown*
>> When will this happen? In general scenario it should be normal that is
>> what i got from internet, but the state is not getting updated in the lease
>> file.
>>
>> On Sat, 5 Oct 2019, 21:16 Bill Shirley, <bill at c3po.polymerindustries.biz>
>> wrote:
>>
>>> Assuming you're referring to DHCP failover, is there any traffic flow on
>>> the
>>> port and peer port in the failover stanza?
>>>
>>> What is your value for mclt?
>>>
>>> Which server, primary or secondary, is giving the recovering message?
>>>
>>> Bill
>>> On 10/5/2019 9:33 AM, Surya Teja wrote:
>>>
>>> Hi I have an issue in the lease flow with isc dhcp service. In the logs
>>> it is printing  * eth0: not responding (recovering) *
>>> My local is set up with active-active mode(splt value as 50-50%) and
>>> because of some reason one of the appliance  went down for some duration. I
>>> observed this and i bring it up, and duration of down is nearly 15hr.
>>> After i bring it up. I am seeing the logs saying not responding
>>> (recovering). Its been more than two hours still I am getting the same
>>> logs
>>> Does any one have any idea about this scenario and how to get the
>>> environment stable
>>>
>>> _______________________________________________
>>> dhcp-users mailing listdhcp-users at lists.isc.orghttps://lists.isc.org/mailman/listinfo/dhcp-users
>>>
>>> _______________________________________________
>>> dhcp-users mailing list
>>> dhcp-users at lists.isc.org
>>> https://lists.isc.org/mailman/listinfo/dhcp-users
>>>
>>
> _______________________________________________
> dhcp-users mailing listdhcp-users at lists.isc.orghttps://lists.isc.org/mailman/listinfo/dhcp-users
>
> _______________________________________________
> dhcp-users mailing list
> dhcp-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20191010/e9a32463/attachment-0001.htm>