[1235] in Hesiod
Re: [Hesiod] Hesiod and long group records
daemon@ATHENA.MIT.EDU (Richard Feltstykket)
Wed Sep 5 01:26:46 2018
Date: Tue, 4 Sep 2018 22:25:54 -0700
From: Richard Feltstykket <rafeltstykket@ucdavis.edu>
To: Win Treese <treese@mit.edu>
Message-ID: <20180905052554.GF31892@odinseye.genomecenter.ucdavis.edu>
MIME-Version: 1.0
In-Reply-To: <200D68E7-A1BE-4744-A416-C497A305C468@mit.edu>
Cc: hesiod@mit.edu
Content-Type: multipart/mixed; boundary="===============3498179931091177992=="
Errors-To: hesiod-bounces@mit.edu
--===============3498179931091177992==
Content-Type: multipart/signed; micalg=pgp-sha1;
protocol="application/pgp-signature"; boundary="0hHDr/TIsw4o3iPK"
Content-Disposition: inline
--0hHDr/TIsw4o3iPK
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
Hi Win,
Thanks for the help again...=20
classes=3DIN,HS is what we have in /etc/hesiod.conf. I tried with just cla=
sses=3DIN, but no change.
Other relevant information I've figured out so far:
1) We have other groups split up in this same manner in DNS which work fine=
, but this is the largest. Once I remove enough members from the workshop.=
group.ns record to get the character count to 512 or less, it starts workin=
g.
2) Problem is present on Ubuntu 14.04 hosts as well.
3) I just discovered that the larger group works fine on our freebsd filese=
rvers:
root@deborah:~ # getent group workshop
workshop:x:2029:fugon,dshill,rpnagarajan,tkremin,dalhindi,cprator,niyer,lix=
inzhe,rbellone,abzzram,paulae,sefeeney,yyaqiaoo,oanh,yiklim,joshi,skinnerjs=
,mclewis,jltanaka,dlebron,malbrown,xycheng,jli,shiqizh,kroy,bpdurbin,emilyk=
,csbenton,sharifdvm,bridget,nikj,mltu,kjennings,alaahmed,rkallen,lifswaro,d=
atawrangler,jennifermoussa,ruta,jaime,markjs,zephyryi,ghanna,jbmiesfeld,sha=
yne,cjjacob,mquesta,limchen,ajmichael,sfyanw,carfranc,agingrich,kluepdan,mi=
nyuan,arios,yangx,msettles,adwright,pmhenry,acartwright,kmlewald,devanmurph=
y,loxiles,yuabai,hforbes,onoses,eokello,ghwada,jfass,jabaily,cincinhuang,ya=
oluo,stafford,sleepery,ragrahn,karin,rf,mbritton,hergarden,cpinzari,ebattiv=
elli,vwang
Thanks,
Richard
On Tue, Sep 04, 2018 at 10:10:21PM -0400, Win Treese wrote:
>=20
> Hi, Richard.
>=20
> I took a quick look at the code. Unfortunately, I don=E2=80=99t have a re=
levant working environment to poke at things.
>=20
> The call to res_mkquery is just using the buffers to construct the query,=
not to handle the response. The array abuf will hold the response, and it =
seems to be big enough (1024 bytes) in both the hesinfo library code and gl=
ibc.
>=20
> Looking at glibc/hesiod/hesiod.c, I=E2=80=99d guess that the exit status =
of 2 from getent that you saw is coming from the __set_errno to ENOENT, whi=
ch is 2. This suggests that there is something the code doesn=E2=80=99t lik=
e about the response. The most likely thing is that the class or type doesn=
=E2=80=99t match, since that seems to be the only clear path in the code to=
ENOENT.
>=20
> Does the configuration of the classes to be used in /etc/hesiod.conf matc=
h the class (IN or HS) being used in your DNS server? If not, that=E2=80=99=
s likely the problem.
>=20
> Best,
>=20
> Win Treese
> treese@acm.org / treese@mit.edu
>=20
> > On Sep 4, 2018, at 6:24 PM, Richard Feltstykket <rafeltstykket@ucdavis.=
edu> wrote:
> >=20
> > Hi List,
> >=20
> > We are running into a problem with large groups on ubuntu 16.04. Hesin=
fo returns the correct record but getent does not.
> >=20
> > rf@odinseye:~$ hesinfo workshop group
> > workshop:x:2029:fugon,dshill,rpnagarajan,tkremin,dalhindi,cprator,niyer=
,lixinzhe,rbellone,abzzram,paulae,sefeeney,yyaqiaoo,oanh,yiklim,joshi,skinn=
erjs,mclewis,jltanaka,dlebron,malbrown,xycheng,jli,shiqizh,kroy,bpdurbin,em=
ilyk,csbenton,sharifdvm,bridget,nikj,mltu,kjennings,alaahmed,rkallen,lifswa=
ro,datawrangler,jennifermoussa,ruta,jaime,markjs,zephyryi,ghanna,jbmiesfeld=
,shayne,cjjacob,mquesta,limchen,ajmichael,sfyanw,carfranc,agingrich,kluepda=
n,minyuan,arios,yangx,msettles,adwright,pmhenry,acartwright,kmlewald,devanm=
urphy,loxiles,yuabai,hforbes,onoses,eokello,ghwada,jfass,jabaily,cincinhuan=
g,yaoluo,stafford,sleepery,ragrahn,karin,rf,mbritton,hergarden,cpinzari,eba=
ttivelli,vwang
> >=20
> > rf@odinseye:~$ getent group workshop
> > rf@odinseye:~$ echo $?
> > 2
> >=20
> > The workshop.group.ns record is in in a single TXT record in DNS like t=
his:
> > "workshop:x:2029:fugon,dshill,rpnagarajan,tkremin,dalhindi,cprator,niye=
r,lixinzhe,rbellone,abzzram,paulae,sefeeney,yyaqiaoo,oanh,yiklim,joshi,skin=
nerjs,mclewis,jltanaka,dlebron,malbrown,xycheng,jli,shiqizh,kroy,bpdurbin,e=
milyk,csbenton,sharifdvm,bridg" "et,nikj,mltu,kjennings,alaahmed,rkallen,li=
fswaro,datawrangler,jennifermoussa,ruta,jaime,markjs,zephyryi,ghanna,jbmies=
feld,shayne,cjjacob,mquesta,limchen,ajmichael,sfyanw,carfranc,agingrich,klu=
epdan,minyuan,arios,yangx,msettles,adwright,pmhenry,acartw" "right,kmlewald=
,devanmurphy,loxiles,yuabai,hforbes,onoses,eokello,ghwada,jfass,jabaily,cin=
cinhuang,yaoluo,stafford,sleepery,ragrahn,karin,rf,mbritton,hergarden,cpinz=
ari,ebattivelli,vwang"
> >=20
> > I was perusing the nss_hesiod directory in the glibc (2.23) source pack=
age for ubuntu 16.04 and also the source for the utility hesinfo and it loo=
ks like the version in glibc may not be new enough to support responses lar=
ger than 1024 bytes. Does that sound accurate?
> >=20
> > In hesiod.c for the hesinfo program I see:
> >=20
> > /* Construct the query. */
> > n =3D res_mkquery(QUERY, name, C_IN, T_TXT, NULL, 0, NULL, qbuf, PACKET=
SZ);
> >=20
> > Whereas in nss_hesiod I see:
> > /*
> > * Construct the query and send it.
> > */
> > n =3D res_mkquery(QUERY, name, class, T_TXT, NULL, 0, NULL, qbuf, MAX_=
HESRESP);
> >=20
> > Am I on the right track here?
> >=20
> > Thanks!
> > Richard
> > _______________________________________________
> > Hesiod@mit.edu
> > http://mailman.mit.edu/mailman/listinfo/hesiod
>=20
--0hHDr/TIsw4o3iPK
Content-Type: application/pgp-signature; name="signature.asc"
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAEBAgAGBQJbj2jiAAoJEJO81r3fDHGU+DoH/1X44NTfYCWlJ6jAfdx7rJKe
p5SqwiFFOEWPKYLTgAWmLPuu2wXzv6XwoWlPijk6ycW5WZR7fAa7xdSewt95RTbX
4ozvWdJHLktY1eeTJ5OLVfvvW+kXgteDso/iQfIdtIoIkwykdg/AjFBN4ol34F99
jmO3tEDHl8A9lraV3C2U5/eTCA+VWw+TihheWEOe87GbaFLpWWBvRdjcfhpDxu2p
SjivdPv0+T3qZ6HxjWVtGgvzFHkMQzntJGDJrvrl/mkOHW7RkhaEWIfRqi8guNLZ
aYZs0S6uUJFY4U7OYP7zweYdAe6WuCVwmX8QP5nWtHoE20oM2eJgtbAGCztOJfo=
=Hm8J
-----END PGP SIGNATURE-----
--0hHDr/TIsw4o3iPK--
--===============3498179931091177992==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
_______________________________________________
Hesiod@mit.edu
http://mailman.mit.edu/mailman/listinfo/hesiod
--===============3498179931091177992==--
