samba - Jul 2002 - Slowdown and errors

Hi all,

   I am having some strange problems on a customer site :

   I am pretty sure I disabled oplocks entirely - I don't like the idea of 
them, and would rather have stable data.

   The customer (5-10 users) runs Sage Line 50, and until recently was 
getting numerous errors within sage.  We have temporarily replaced their 
switch and this has reduced the fatality rate and a few corruptions.  They 
still however get a message along the lines of "waiting for file 
index.dta".  I have looked into the logs and found no errors apart from the
standard anonymous access not allowed type messages.  If anyone knows 
wether Sage Line 50 requires oplocks or has a configuration that works 
solidly with it, please step forward.  I am awaiting their connection to 
the net to retrieve their smb.conf and testparm output, but anything to 
look for specifically would be a boon.

   Thanks in advance,


Martyn Ranyard
Free Software Advocate

jabber - joran@amessage.de
icq    - 122500800
irc    - Joran on oftc
msn    - ranyardm@hotmail.com
y!     - ranyardm
e      - ranyardm@lineone.net

Why did you have to replace the switch? This could be entirely hardware
related, and nothing to do with software. I've seen customer's sites
before
that worked great, and one day, boom nothing worked right anymore. After
replacing LOTS of hardware, everything worked great again. If you had the
replace the switch because of something electrical, I'd start replacing
NICs too and see if that makes improvements.

At 11:10 AM 7/5/2002 +0100, you wrote:
>
>Hi all,
>
>   I am having some strange problems on a customer site :
>
>   I am pretty sure I disabled oplocks entirely - I don't like the idea
of
>them, and would rather have stable data.
>
>   The customer (5-10 users) runs Sage Line 50, and until recently was 
>getting numerous errors within sage.  We have temporarily replaced their 
>switch and this has reduced the fatality rate and a few corruptions.  They 
>still however get a message along the lines of "waiting for file 
>index.dta".  I have looked into the logs and found no errors apart from
the
>standard anonymous access not allowed type messages.  If anyone knows 
>wether Sage Line 50 requires oplocks or has a configuration that works 
>solidly with it, please step forward.  I am awaiting their connection to 
>the net to retrieve their smb.conf and testparm output, but anything to 
>look for specifically would be a boon.
>
>   Thanks in advance,
>
>
>Martyn Ranyard
>Free Software Advocate
>
>jabber - joran@amessage.de
>icq    - 122500800
>irc    - Joran on oftc
>msn    - ranyardm@hotmail.com
>y!     - ranyardm
>e      - ranyardm@lineone.net
>
>
>-- 
>To unsubscribe from this list go to the following URL and read the
>instructions:  http://lists.samba.org/mailman/listinfo/samba
>

On Fri, Jul 05, 2002 at 11:10:03AM +0100, Martyn Ranyard
wrote:
> 
> Hi all,
> 
>    I am having some strange problems on a customer site :
> 
>    I am pretty sure I disabled oplocks entirely - I don't like the idea
of
> them, and would rather have stable data.
> 
>    The customer (5-10 users) runs Sage Line 50, and until recently was 
> getting numerous errors within sage.  We have temporarily replaced their 
> switch and this has reduced the fatality rate and a few corruptions.  They 
> still however get a message along the lines of "waiting for file 
> index.dta".  I have looked into the logs and found no errors apart
from the
> standard anonymous access not allowed type messages.  If anyone knows 
> wether Sage Line 50 requires oplocks or has a configuration that works 
> solidly with it, please step forward.  I am awaiting their connection to 
> the net to retrieve their smb.conf and testparm output, but anything to 
> look for specifically would be a boon.
Someone working with a FoxPro database found that they had to
increase the number of attempts to get a byte range lock in
order to make it work correctly. This is an error in the
application but Windows NT/2000/XP (and now Samba) works
around it by retying lock attempts.

The two parameters you could try increasing are :

lock spin count
lock spin time

which are currently set to "lock spin count = 3",
"lock spin time = 10 (usec)". I can't remember the
exact value (although I can look them up) but I think
it was a large increase in the lock spin count to
something like 20 instead of 3 to make the application
work right.

This could be interesting to help correctly set the
default values for these parameters so if you could
let the list know if changing them helps.

Thanks,

	Jeremy.

Hi Martyn

We have been using Sage50 (Windows 98 clients) and Sage 100 (Windows NT/2000
clients) successfully since Nov 2001 on Samba 2.2.x.  Oplocks are off of
course and we have nothing fancy in our smb.conf (below).

The only problems we have had with Sage are when the users neglect to clear
out their print spool directory (Sage creates three files for each print).
When they browse their printout listings Samba goes berserk as it tries to
index directories with thousands of files in them (load average 8+).  Novell
never blinked on the same hardware.  As long as these spool directories are
tidied regularly then this is avoided.

You might want to examine the customer's network infrastructure a little
more if they had a duff switch?  They could have other nasties in there.

HTH,
Cheers,
Noel

[global]
        workgroup = UK
        netbios name = BELLY
        server string = Samba Server 2.2.4
        invalid users = root bin uucp sys
        encrypt passwords = Yes
        update encrypted = Yes
        keepalive = 30
        deadtime = 5
        socket options = IPTOS_LOWDELAY TCP_NODELAY
        host msdfs = yes
        utmp = yes

        os level = 0
        preferred master = False
        local master = No
        domain master = False

        security = domain
        password server = BRAIN,MOUTH
        smb passwd file = /usr/local/samba/private/smbpasswd

        wins server = 192.168.5.4
        name resolve order = wins host bcast

        winbind uid = 10000-20000
        winbind gid = 10000-20000
        winbind enum users = yes
        winbind enum groups = yes
        winbind separator = +
        nt acl support = yes
        template homedir = /raid/homedrives/%U
        template shell = /dev/null

        # These oplock settings increase file access dramatically but
        # we might have to negate them if we experience run away smbd
processes
        kernel oplocks = no
        oplocks = no
        level2 oplocks = no

        #panic action = "/root/panic.sh %d"
        #panic action = /usr/bin/X11/xterm -display :2.0 -e gdb
/usr/local/samba
/bin/winbindd
#       log file = /usr/local/samba/var/log.%m
        log level = 1


[SAGE]
        path = /raid/shared/SAGE/
        browseable = no
        public = no
        read only = No
        inherit permissions = yes
        create mask = 777
        directory security mask = 777
        nt acl support = yes
        valid users = @uk+it,@uk+accounts,@uk+finance

>    The customer (5-10 users) runs Sage Line 50, and until recently was 
> getting numerous errors within sage.  We have temporarily replaced their 
> switch and this has reduced the fatality rate and a few corruptions.  They
> still however get a message along the lines of "waiting for file 
> index.dta".  I have looked into the logs and found no errors apart
from
the 
> standard anonymous access not allowed type messages.  If anyone knows 
> wether Sage Line 50 requires oplocks or has a configuration that works 
> solidly with it, please step forward.  I am awaiting their connection to 
> the net to retrieve their smb.conf and testparm output, but anything to 
> look for specifically would be a boon.

---
Outgoing mail is certified Virus Free.
Checked by AVG anti-virus system (http://www.grisoft.com).
Version: 6.0.373 / Virus Database: 208 - Release Date: 01/07/2002