Bug#42097: libc6: "clnt_call: RPC: Timed out" in nis
In article <cistron.E119cEJ-0003My-00@chinook.stanford.edu>,
<maxk@chinook.stanford.edu> wrote:
>Package: libc6
>Version: 2.1.2-0pre2
>Severity: normal
>
>I am using nis 3.3.3-1 and I'm getting the following error when trying
>to log in or doing an su on nis client machines:
>
>yp_all: clnt_call: RPC: Timed out
>
>According to the maintainer of the nis package, this is a glibc bug
>and not something in the nis package. Both the server and client are
>running the same versions of all the packages.
I talked to Thorsten, the guy who maintains both the glibc NIS code
and the NIS server / utilities, and I asked him whether this bug is
in the NIS tools or in glibc:
On Thu, Jul 29, Miquel van Smoorenburg wrote:
> According to Thorsten Kukuk:
> > You should consider to use ypserv 1.3.6.94 (or 1.3.7 if I find the time
> > the next days to release it).
> > There was a very bad bug in yp_all(). As result clients could get wrong data
> > back. It nearly never happens with libc5/glibc 2.0, but glibc 2.1 uses
+yp_all ()
> > in much more places.
>
> Thorsten,
>
> I got a few compliants like this:
>
> > After upgrading to the new 3.3.3 version of nis in potato, I started > > getting the following errors when logging in or doing an su on nis
> > client machines:
> >
> > yp_all: clnt_call: RPC: Timed out
>
> Is that the error you are talking about?
Since you get it on logging in or doing su, it could be. But I don´t
know which glibc version you are using, there was a bug in sunrpc
from glibc 2.1 which shows the same result. glibc 2.1.1 should be Ok.
The sunrpc bug also causes ypbind-mt and ypserv to end in a loop
eating 100% CPU time. Instead of telling it to me or fixing the
bug, RH only says ypbind-mt sucks and removed it :(
He will make an ypserv-1.3.7 release soon he promised, and I'll
create a new NIS package (3.4) with ypserv-1.3.7 as soon as possible
after that. Please see if that helps, if it does it was an ypserv
bug and this bug report can be closed. If it doesn't, then I
really suspect this still is a glibc problem.
Mike.
--
... somehow I have a feeling the hurting hasn't even begun yet
-- Bill, "The Terrible Thunderlizards"
Reply to: