[Samba] Re: [fwd: [Fwd: Re: network response] ]

[Jeremy Allison]

On Thu, Jul 01, 2004 at 10:53:03AM -0700, Jeremy Allison wrote:
> On Thu, Jul 01, 2004 at 11:05:18AM -0500, Chris Garrigues wrote:
> > > From:  Chris Garrigues <cwg-dated-1089032093.48bcf5 at deepeddy.com>
> > > Date:  Wed, 30 Jun 2004 07:54:53 -0500
> > >
> > > (I'm still trying to get a new debug log for you.)
> > 
> > Sorry this took a while, but unfortunately I have other responsibilities.
> > 
> > New log at
> > 
> > 	http://www.trinsics.com/~cwg/smblog
> > 
> > The delay was from 10:09:52 to 10:11:17.
> 
> At last ! Thanks !!!!!
> 
> The problem is here :
> 
> 249646 [2004/07/01 10:09:52, 5] lib/smbldap.c:smbldap_search(932)  249647   smbldap_search: base => [dc=trinsics,dc=com], filter => [(&(sambaSID=S-1-5-21-904170827-898255731-2295272421-1407)(        objectclass=sambaSamAccount))], scope => [2]
>  249648 [2004/07/01 10:09:52, 5] lib/smbldap.c:smbldap_search(950)
>  249649   smbldap_search: waiting 1117501 milliseconds for LDAP replication.
>  249650 [2004/07/01 10:11:17, 6] param/loadparm.c:lp_file_list_changed(2687)
>  249651   lp_file_list_changed()
>  249652   file /etc/samba3/smb.conf -> /etc/samba3/smb.conf  last mod_time: Wed Jun 30 16:11:03 2004
>  249653
>  249654 [2004/07/01 10:11:17, 5] smbd/connection.c:claim_connection(170)
>  249655   claiming  0
>  249656 [2004/07/01 10:11:17, 0] lib/util_sock.c:get_peer_addr(1000)
>  249657   getpeername failed. Error was Transport endpoint is not connected
> 
> 
> Nothing to do with the new open code. There is a bug (not sure
> what yet) with calculating the sleep time in smbldap_search()
> (in fact I have *NO* idea what this code is for....) that causes
> it to sleep for much more than 30 second - client times out, fun
> ensues....

Ok, I've fixed up the smbldap_search code and added a max sleep replication
time of 5 seconds. I've committed to 3.0 svn if you want to try it. This
should fix the issue you've reported here. Let me know if not....

Thanks,

	Jeremy.