Hi Eric;
HP didn't put Samba/iX on 5.5, therefore NONE of the versions is 'tested'
on 5.5 (by the hp labs anyway...) So there aren't any patches for Samba/iX
on 5.5 either.
But if you were having the log rollover problem, what that actually ends up
doing
is 'hanging' smbd and nmbd, so that they are no longer pulling messages and
dealing with them, so you end up running out of network buffers in general
on MPE/iX,
and everything starts going south... So this is probably NOT what is
happening to you.
From your first message it sounded like everything worked fine for a certain
period of
time (hours up to weeks sometimes), and then 'certain' pc's would no longer
be able to
connect to samba, but other pc's would still be working. And the pc's
affected are not
the same each time.
this might be the case if what is REALLY going on is that the pc's that are
STILL working
never lost their connection to SAMBA/iX, so they already(still) have a
working smbd process.
The pc's that are NOT working have possibly lost their connection sometime
in the past
(due to no traffic, auto disconnect, whatever), and now when Windows tries
to 'transparently'
reconnect to the server//share in the background, the reconnection is
failing because of
what you are seeing in the log file:
Added interface ip=<system ip> bcast=<bcast addr> nmask=<nmask>
> 07/06/01 15:05:12 loaded services
> standard input is not a socket, assuming -D option
> Unsupported Samba/iX 0.7 by [log in to unmask] 4/97
> 07/06/01 15:05:12 staying in job mode on MPE
> bind failed on port 139 socket_addr=0 (Address already in use)
What that is telling me is that SOME process/service/job is bound to port
137 on the ipaddress
that the pc's that are having the problem are trying to attach to, and a new
smbd process cannot
be spawned because it can't bind to port 139 (the session port).
So that's what you have to figure out at this point; I would still recommend
that until you get
this issue resolved that you remove smbd and nmbd from your inetd.conf file,
bounce inetd, and
then start smbd and nmbd up via the jnmbd and jsmbd job files. THEN if this
problem occurs again,
before killing the jobs and restarting them, work with the customer to look
at all the open ports
and connections involving port 139 on the system using nettool.net.sys
(SOCKINFO)and figure out
who is hogging that port...
Sorry, there may be a better way to do this - I'm not as up on MPE/iX
networking as I should be..
Don
-----Original Message-----
From: Jansen, Eric [mailto:[log in to unmask]]
Sent: Monday, July 09, 2001 5:04 PM
To: 'MCCALL,DON (HP-USA,ex1)'; [log in to unmask];
[log in to unmask]
Subject: RE: samba unreachable from some pc's and not from others
Don,
setting the log-level to 0 didn't help.
Do you know what version of Samba is tested on 5.5 and if the hanging is
resolved in that or is it a patch at the system level?
Kind regards,
Eric Jansen
Technical Consultant
Computer Design & Integration, LLC
696 Route 46 West, Teterboro, NJ 07608
Phone: 201-931-1420 x313
Fax: 201-931-0101
Pager: 973-205-1236
[log in to unmask]
www.cdillc.com
> -----Original Message-----
> From: MCCALL,DON (HP-USA,ex1) [mailto:[log in to unmask]]
> Sent: Monday, July 09, 2001 10:05 AM
> To: 'Jansen, Eric'; [log in to unmask]; [log in to unmask]
> Subject: RE: samba unreachable from some pc's and not from others
>
>
> Hello Jansen,
> On MPE/iX I would recommend that you do NOT run samba
> under inetd control; stream the job files instead. I have heard some
> instances where services run under inetd would hang (ftp included,
> sometimes), but I don't have any specifics.
> There is also an issue with Samba/iX where the log files
> would not roll over
> in the old 5.5 version. Make sure that in your smb.conf file
> you have log
> level = 0 and log size = 10000, so you don't hit a point
> where samba needs
> to rename the logs, and hangs from that. (That issue is
> fixed in patches
> for Samba/iX for 6.0, 6.5 & 7.0 MPE/iX).
>
> Hope this helps,
> Don
>
> -----Original Message-----
> From: Jansen, Eric [mailto:[log in to unmask]]
> Sent: Monday, July 09, 2001 9:50 AM
> To: [log in to unmask]; [log in to unmask]
> Subject: samba unreachable from some pc's and not from others
>
>
> Hi all,
>
> One of our clients has samba version 1.9.16p9 running on a
> HP3000 MPE system
> with OS 5.5 05 p9 (yes, I know it is an old one but the
> client doesn't want
> to upgrade).
>
> After some time of working fine; sometimes weeks, sometimes
> hours; samba is
> unreachable from some pc's.
> We don't really see a certain scenario in which it happens
> neither do we
> have problems with certain pc's, they differ every time.
>
>
> log.smb error:
> -------------
> Added interface ip=<system ip> bcast=<bcast addr> nmask=<nmask>
> 07/06/01 15:05:12 loaded services
> standard input is not a socket, assuming -D option
> Unsupported Samba/iX 0.7 by [log in to unmask] 4/97
> 07/06/01 15:05:12 staying in job mode on MPE
> bind failed on port 139 socket_addr=0 (Address already in use)
> -------------
>
> I checked the network settings with a CISCO network
> specialist and all was
> fine. We also "sniffered" the traffic to the HP3000 and have seen the
> following errors: "window frozen" and "acknowledgement too long". This
> usually reflects to a time-out. Since pinging the system gave a quick
> response time, the network is fine (right?).
>
> To me it seems that the port 139 got hung up for a reason.
>
> Has anybody experienced the same and what can I do to solve it?
> For now we bounce the inetd.
>
> Thanks in advance for your help.
>
> Kind regards,
>
>
> Eric Jansen
> --
> To unsubscribe from this list go to the following URL and read the
> instructions: http://lists.samba.org/mailman/listinfo/samba
>
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
