[Kea-users] problem with Postgres leases after upgrade (I think)

[Tim Howe]

On Tue, 30 Apr 2019 17:35:29 -0700
Tim Howe <tim.h at bendtel.com> wrote:

> I had a working kea 1.3 server.  After upgrading to 1.5 and upgrading
> my PostgreSQL schema, it appears as if Kea isn't able to correctly use
> the new Pg schema(?)
> 
> The below errors seem to suggest that it doesn't like the data that is
> in the hwtype column of the lease6 table.  In the database all that
> data is blank (null?).
> 
> Is there some default value I can fill in that will allow it to get past this problem?
> 
> 
> 
> 2019-04-30 17:12:11.800 DEBUG [kea-dhcp6.leases/10281] DHCP6_PROCESS_IA_NA_REQUEST duid=[00:03:00:01:c8:54:4b:54:e5:28], tid=0xcf349e: server is processing IA_NA option with iaid=1 and hint=(no hint)
> 2019-04-30 17:12:11.800 DEBUG [kea-dhcp6.dhcpsrv/10281] DHCPSRV_PGSQL_GET_IAID_DUID obtaining IPv4 leases for IAID 1 and DUID 00:03:00:01:c8:54:4b:54:e5:28, lease type 0
> 2019-04-30 17:12:11.801 ERROR [kea-dhcp6.alloc-engine/10281] ALLOC_ENGINE_V6_ALLOC_ERROR duid=[00:03:00:01:c8:54:4b:54:e5:28], tid=0xcf349e: error during attempt to allocate an IPv6 address: Could not convert data to Lease6, reason: Invalid data:[] for row: 0 col: 13,hwtype : bad lexical cast: source type value could not be interpreted as target
> 2019-04-30 17:12:11.801 DEBUG [kea-dhcp6.leases/10281] DHCP6_LEASE_ADVERT_FAIL duid=[00:03:00:01:c8:54:4b:54:e5:28], tid=0xcf349e: failed to advertise an address lease for iaid=1
> 2019-04-30 17:12:11.801 DEBUG [kea-dhcp6.options/10281] DHCP6_ADD_STATUS_CODE_FOR_IA duid=[00:03:00:01:c8:54:4b:54:e5:28], tid=0xcf349e: adding Status Code to IA with iaid=1: NoAddrsAvail(2) "Sorry, no address could be allocated."
> 
> --TimH
> _______________________________________________
> Kea-users mailing list
> Kea-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/kea-users


	I think the problem is in src/lib/dhcpsrv/pgsql_lease_mgr.cc

	I notice the state column is ignored here (around line 805).
Is my data different after schema upgrade than would be expected if I
started with 1.5?  Seems this is a data sanity check problem, but I
haven't looked at this code before.

--TimH