--- On Fri, 1/21/11, Andrew Deason <[email protected]> wrote:

> From: Andrew Deason <[email protected]>
> Subject: [OpenAFS] Re: PAG lost after switching users on Slackware 13.1 box
> To: [email protected]
> Date: Friday, January 21, 2011, 3:53 PM
> On Thu, 20 Jan 2011 18:17:00 -0800
> (PST)
> Tom Mukunnemkeril <[email protected]>
> wrote:
> 
> > Is there anything I can do to debug this further to
> figure out why the
> > PAG is getting lost?  The version of glibc is
> different, and I don't
> > know if this would cause a problem, or is there
> something I need to
> > enable in the kernel.
> 
> You can look at what what is happening to your PAG, which
> can be
> identified by the output of 'id -a' and/or 'keyctl show'.
> Try running
> these once you obtain a PAG and after you lose it. At first
> you should
> see something like:
> 
> # id -a
> uid=0(root) gid=0(root) groups=0(root),1107203568
> 
> and
> 
> # keyctl show
> Session Keyring
>        -3 --alswrv   
>   0     0  keyring:
> _ses.21944
> 271956261 ----s--v      0 
>    0   \_ afs_pag: _pag
> 
> The last big group in the 'id' output, and the afs_pag key
> in your
> session keyring can be used to identify your PAG. After you
> su, I'd
> expect both of them to be gone. Is that the case?
> 

id shows a value, but /proc/keys does not.  In the Slackware 13.0 installation, 
/proc/keys does show a value after I execute pagsh.


bash-4.1# pagsh
sh-4.1# id
uid=0(root) gid=0(root) 
groups=0(root),1(bin),2(daemon),3(sys),4(adm),6(disk),10(wheel),1094319331
sh-4.1# cat /proc/keys
sh-4.1# 



> It may have something to do with how much su sanitizes the
> environment,
> but I'm not sure.
> 
> > Both Slackware distributions run Openafs-1.4.12.1 and
> Kernel
> > 2.6.35.7.
> 
> You may want to use OpenAFS 1.4.14. There's at least one
> change in there
> for 2.6.35, though it shouldn't be affecting this.
> 
> -- 

I will try to upgrade OpenAFS 1.4.14 but have doubts this will help since my 
Slackware 13.0 installation with linux 2.6.35.4 and OpenAFS-1.4.12.1 seems to 
work fine.


Tom



_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info

Reply via email to