[avahi] No slot available for legacy unicast reflection

Lennart Poettering lennart at poettering.net
Thu Mar 27 06:16:17 PDT 2008


On Thu, 27.03.08 01:14, Tannahill, Jim (Jim.Tannahill at disney.com) wrote:

> Running Avahi 0.6.22 on Suse Enterprise 10.
> 
> We are using the avahi-daemon mDNS reflection capability and it's
> working great for us across three subnets. However every day we get over
> a million lines of this message:
> 
> "No slot available for legacy unicast reflection, dropping query
> packet."  
> 
> The only place I could find anything about this message was in the code.
> Can someone fill me in on what causes the message and what we need to do
> about it other than rotate the message log a lot . . .

Hmm, someone is apparently using a lot of legacy unicast queries. How
big is your network?

Try increasing AVAHI_LEGACY_UNICAST_REFLECT_SLOTS_MAX to 1000 in
avahi-core/internal.h. This will allocate another 900 slots for
storing information about individual legacy unicast
queries. Apparently your clients use more than 100 of them. They
usually time out after 2s.

If you want to know what legacy unicast queries actually are, check
section 8.3 of the mdns spec. Basically they are traditional DNS
queries sent to the multicast group, in contrast to the modernized
native mDNS queries that are normally sent to the mcast group.

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553
http://0pointer.net/lennart/           GnuPG 0x1A015CC4


More information about the avahi mailing list