Je bind-4.9.7-1 broken?

Pavel Kankovsky peak na argo.troja.mff.cuni.cz
Čtvrtek Květen 20 21:17:12 CEST 1999


On Thu, 20 May 1999, Petr Novotny wrote:

> pouzivam bind-4.9.7-1 (to je ten z updated pro RedHat 5.1). Mam 
> pocit, ze nedela TCP-retry v pripade, ze UDP odpoved je moc 
> dlouha. Muzete toto chovani nekdo potvrdit/vyvratit? (Zkuste se 
> treba pres nslookup podivat na MX pro large-mx.ckdhr.com - 
> odpoved od vaseho lokalniho nameserveru by mela byt stejna jako 
> odpoved od nejakeho autoritativniho, treba ns.chicago.com.)

hmm... skutecne se to tak chova (pricemz pravdepodobnost, ze ma maslo na
hlave firewall nebo nejaka RH uprava je mala)

mozna to nebude nahoda:

(citat z ns_resp.c)

        if (hp->tc) {
                count -= arcount;       /* truncation had to affect this */
                if (!arcount) {
                        count -= aucount;       /* guess it got this too */
                }
                if (!(arcount || aucount)) {
                        count -= ancount;       /* things are pretty grim */
                }
                /* XXX - should retry this query with TCP */
                /*
                 * XXX - if this response is forwarded to the client
                 * the truncated section is included.  We will not
                 * validate it, and if it somehow corrupt, we won't
                 * notice.
                 *
                 * XXX - if the answer section is truncated and we got
                 * this response after being redirected by a CNAME, we
                 * will not include any part of the final answer in our
                 * response to the client.  This will make the client
                 * think that there are no RRs of the appropriate type.
                 */
        }

:P

--Pavel Kankovsky aka Peak  [ Boycott Microsoft--http://www.vcnet.com/bms ]
"NSA GCHQ KGB CIA nuclear conspiration war weapon spy agent... Hi Echelon!"



Další informace o konferenci Linux