@lists.samba.org
Subject: Re: [Samba] Very slow samba performance on Centos 6
On Thu, Aug 04, 2011 at 10:49:50AM -0400, vg_ us wrote:
I have 2 identical Dell r510 servers with 10gig card, running centos
6 with samba-3.5.4-68.el6_0.2.x86_64.
I setup 16G ramdisk samba share on both and ran cp
,
Robert Adkins II
-Original Message-
From: samba-boun...@lists.samba.org
[mailto:samba-boun...@lists.samba.org] On Behalf Of vg_ us
Sent: Thursday, August 04, 2011 2:12 PM
To: volker.lende...@sernet.de
Cc: samba@lists.samba.org
Subject: Re: [Samba] Very slow samba performance
-
From: Robert Adkins II radk...@impelind.com
Sent: Friday, August 05, 2011 10:10 AM
To: 'vg_ us' vg...@hotmail.com; volker.lende...@sernet.de
Cc: samba@lists.samba.org
Subject: RE: [Samba] Very slow samba performance on Centos 6
Wouldn't
On Thu, Aug 04, 2011 at 10:49:50AM -0400, vg_ us wrote:
I have 2 identical Dell r510 servers with 10gig card, running centos
6 with samba-3.5.4-68.el6_0.2.x86_64.
I setup 16G ramdisk samba share on both and ran cp from local
ramdisk to samba ramdisk mount.
If I cp 12 1-gig files, I get
--
From: Volker Lendecke volker.lende...@sernet.de
Sent: Thursday, August 04, 2011 11:01 AM
To: vg_ us vg...@hotmail.com
Cc: samba@lists.samba.org
Subject: Re: [Samba] Very slow samba performance on Centos 6
On Thu, Aug 04, 2011 at 10:49:50AM
On 8/4/2011 1:11 PM, vg_ us wrote:
cifsfs mounts are really slow, so what happens when linux, windows and
mac clients map/mount the share? Are they gonna be this slow? Any way to
speed it up?
Unfortunately I don't have an answer to the slow mounts issue. However,
you're showing a peak
On Mon, 2011-07-25 at 19:51 -0400, simo wrote:
On Tue, 2011-07-26 at 00:32 +0100, Jonathan Buzzard wrote:
Jeremy Allison wrote:
[SNIP]
Test using a modern (i.e. much later than 3.0.33) smbclient.
To back that up he is using CentOS 5, so there is no excuse for using
These are XP clients.
Date: Mon, 25 Jul 2011 13:28:33 -0700
From: j...@samba.org
To: groucho.64...@hotmail.com
CC: samba@lists.samba.org
Subject: Re: [Samba] Very slow write performance to RAID
On Mon, Jul 25, 2011 at 01:06:48PM -0400, Kevin Taylor wrote:
We have a RAID set up
On Mon, Jul 25, 2011 at 12:06 PM, Kevin Taylor
groucho.64...@hotmail.com wrote:
We have a RAID set up as our main fileserver (running samba 3.0.33 on linux,
CentOS 5). The main disk area is an XFS partition of about 8TB. I'm using
iostat to monitor disk I/O since we've gotten complaints
@lists.samba.org
Subject: Re: [Samba] Very slow write performance to RAID
On Mon, Jul 25, 2011 at 12:06 PM, Kevin Taylor
groucho.64...@hotmail.com wrote:
We have a RAID set up as our main fileserver (running samba 3.0.33 on
linux, CentOS 5). The main disk area is an XFS partition of about
On Mon, Jul 25, 2011 at 01:06:48PM -0400, Kevin Taylor wrote:
We have a RAID set up as our main fileserver (running samba 3.0.33 on linux,
CentOS 5). The main disk area is an XFS partition of about 8TB. I'm using
iostat to monitor disk I/O since we've gotten complaints about speed and I'm
Jeremy Allison wrote:
[SNIP]
Test using a modern (i.e. much later than 3.0.33) smbclient.
To back that up he is using CentOS 5, so there is no excuse for using
such an old version. Needs to switch to the samba3x packages that have
been present since CentOS 5.5 asap. From recollection it
On Tue, 2011-07-26 at 00:32 +0100, Jonathan Buzzard wrote:
Jeremy Allison wrote:
[SNIP]
Test using a modern (i.e. much later than 3.0.33) smbclient.
To back that up he is using CentOS 5, so there is no excuse for using
such an old version. Needs to switch to the samba3x packages
On Wed, Jun 24, 2009 at 12:26:59PM -0700, John Du wrote:
Have you looked at adjusting the socket options parameter in smb.conf?
I use socket options = TCP_NODELAY SO_SNDBUF=8192 SO_RCVBUF=8192 in my
smb.conf and scp and samba take about the same amount of time to
transfer files from
What type of file processors are you running along with samba?. Are you
running the virus checking plugin or VFS(recycle bin)? Virus checking is
very cpu and disk I/O intensive these can really slow down a samba
server. I can't expect VFS is all that cheap either when moving big files.
Raghu A
There is no disk or CPU bottleneck or virus checking (server is latest
ubuntu). scp at the same time as this transfer can write 3-4 faster to the
same partition. This is an Atom processor but there is more cpu left.
To be more specific:
Why does XP send only 1KB at a time to the server? I think
Have you looked at adjusting the socket options parameter in smb.conf?
I use socket options = TCP_NODELAY SO_SNDBUF=8192 SO_RCVBUF=8192 in my
smb.conf and scp and samba take about the same amount of time to
transfer files from Windows to the samba server.
Raghu A wrote:
There is no disk or
I tried TCP_NODELAY and it didn't make a difference. I haven't tried
SO_SNDBUF and RCVBUF, but I will. As the tcpdump shows there is lot of tcp
window left.
It is not just the server since linux samba client behaves much better. What
determines SMB packet size?
What is the throughput you get?
My samba server is 3.0.28a running on RHEL 4.
My network is also 100Mbps. I copy a 100MB file from Windows XP to my
samba server in about 20 seconds. Scp the same file from a Linux host
to the same server takes about the same time.
You may take a look at setting the Windows TCP buffer
Well, 1KB is infact the culprit and it seems to be an artifact of the
application. SMB seems to use one packet for each write() call from the
app. So my cygwin command on XP (cat file remote_file) must be calling
write(1KB). I controlled this write size with dd command and sure enough I
hit the
Thanks Ben and John, for looking into this issue. I want to clarify since my
earlier comment might be misinterpreted to imply I didn't appreciate your
help. I certainly did.
Raghu A wrote:
Well, 1KB is infact the culprit and it seems to be an artifact of the
application. SMB seems to use
Sample tcpdump for such a connection : Notice that there are only couple of
1KB chunks for each millisecond :
18:50:57.948157 IP 192.168.0.100.4366 192.168.0.104.445: P 2184:3276(1092)
ack 103 win 64719
18:50:57.948374 IP 192.168.0.104.445 192.168.0.100.4366: P 103:154(51) ack
3276 win 65535
Jeremy Allison wrote:
On Fri, Aug 29, 2008 at 12:54:53PM -0500, Gregory Carter wrote:
Oh, and BY THE WAY.
I do not want to be a total cynic, but you are expecting samba to
replace a software product that the SuSE corporation directly receives
MILLIONS in contributions from, said vendor of
The newest Samba for RHEL 5.2 should be 3.0.28. Is there a reason this box
isn't up to date?
On Fri, Aug 29, 2008 at 6:16 PM, Brian D. McGrew [EMAIL PROTECTED]wrote:
So now after I've been playing around with the configuration and such, it
seems that the SMB server has become less usable.
Subject: Re: [Samba] Very Slow!
The newest Samba for RHEL 5.2 should be 3.0.28. Is there a reason this box
isn't up to date?
On Fri, Aug 29, 2008 at 6:16 PM, Brian D. McGrew [EMAIL PROTECTED] wrote:
So now after I've been playing around with the configuration and such, it seems
that the SMB
Brian McGrew wrote:
System info:
Red Hat Enterprise Linux Server release 5 (Tikanga)
Kerlen 2.6.18-8.el5 SMP x86_64
Samba version 3.0.23c-2
Eth0 Eht1 bonded to bond0, 2Gbps.
/etc/samba/smb.conf attached below...
I¹m seeing very slow transfers from Samba I¹m not sure
On Fri, 2008-08-29 at 10:49 -0400, Gerald Drouillard wrote:
Brian McGrew wrote:
System info:
Red Hat Enterprise Linux Server release 5 (Tikanga)
Kerlen 2.6.18-8.el5 SMP x86_64
Samba version 3.0.23c-2
Eth0 Eht1 bonded to bond0, 2Gbps.
Try unbonding the NICs and just
I am going to go with a bad samba build.
Won't be the first time.
Try different rpm versions from Red.
Update or Backrev
If that still doesn't work, try putting both the client and the server
on a unmanaged gigabit switch and try the test again.
-gc
Brian McGrew wrote:
System info:
Oh, and BY THE WAY.
I do not want to be a total cynic, but you are expecting samba to
replace a software product that the SuSE corporation directly receives
MILLIONS in contributions from, said vendor of product it is replacing.
(Microsoft.)
SuSe would be the absolute LAST linux distro I
On Fri, Aug 29, 2008 at 12:54:53PM -0500, Gregory Carter wrote:
Oh, and BY THE WAY.
I do not want to be a total cynic, but you are expecting samba to
replace a software product that the SuSE corporation directly receives
MILLIONS in contributions from, said vendor of product it is
On 8/29/2008 1:54 PM, Gregory Carter wrote:
Oh, and BY THE WAY.
I do not want to be a total cynic, but you are expecting samba to
replace a software product that the SuSE corporation directly receives
MILLIONS in contributions from, said vendor of product it is replacing.
(Microsoft.)
On 8/29/08 11:16 AM, Jeremy Allison [EMAIL PROTECTED] wrote:
On Fri, Aug 29, 2008 at 12:54:53PM -0500, Gregory Carter wrote:
Oh, and BY THE WAY.
I do not want to be a total cynic, but you are expecting samba to
replace a software product that the SuSE corporation directly receives
MILLIONS
So now after I've been playing around with the configuration and such, it seems
that the SMB server has become less usable. Now, all the shares are visible
but as soon as I try to access anything or copy anything I get The network
path is not valid. Again, trying from XP, 2003 and 2008.
I've
On Thu, Aug 28, 2008 at 02:20:09PM -0700, Brian McGrew wrote:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192
Quick try: Remove that.
Curious question -- why did you set those options?
Volker
pgpixy5KnCfiL.pgp
Description: PGP signature
--
To unsubscribe from this list go to
On Thu, Aug 28, 2008 at 02:20:09PM -0700, Brian McGrew wrote:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192
Quick try: Remove that.
Curious question -- why did you set those options?
-
It didn't change, still says 4 hours and is taking 3 to 4 seconds to copy
1k.
I
On Thu, Aug 28, 2008 at 02:34:02PM -0700, Brian McGrew wrote:
On Thu, Aug 28, 2008 at 02:20:09PM -0700, Brian McGrew wrote:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192
Quick try: Remove that.
Curious question -- why did you set those options?
-
It didn't
Volker Lendecke wrote:
On Thu, Aug 28, 2008 at 02:20:09PM -0700, Brian McGrew wrote:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192
Quick try: Remove that.
Curious question -- why did you set those options?
Volker
That is in the default smb.conf distributed with many
On Thu, Aug 28, 2008 at 02:34:02PM -0700, Brian McGrew wrote:
On Thu, Aug 28, 2008 at 02:20:09PM -0700, Brian McGrew wrote:
socket options = TCP_NODELAY SO_RCVBUF=8192 SO_SNDBUF=8192
Quick try: Remove that.
Curious question -- why did you set those options?
-
It didn't
You could try setting some of the oplock options. I think fake oplocks =
yes on the
application share could significantly increase the performance when executing.
Éder,
Thanks for your response.
The slow share problem turned out to be a Samba/Ubuntu 7.04 problem with some
NICs -
symptoms
Christian Perrier escribió:
Quoting Eric Shuman ([EMAIL PROTECTED]):
Hi all,
I am having a problem accessing very large files through my samba shares
after upgrading my file server to Debian Etch (Samba 3.0.24) from Debian
Sarge (Samba ???).
Debian sarge has 3.0.14a
I'm
Quoting Eric Shuman ([EMAIL PROTECTED]):
Hi all,
I am having a problem accessing very large files through my samba shares
after upgrading my file server to Debian Etch (Samba 3.0.24) from Debian
Sarge (Samba ???).
Debian sarge has 3.0.14a
--
To unsubscribe from this list go to the
How about the temp directory in Office Preferences. Is it on the local
workstation, or does it default to the dir where the file is opened. I
always thought that could slow things down. (i.e. the creation of the
~foo.doc file).
HTH,
Steve
--
To unsubscribe from this list go to the following
Berend Tober ha scritto:
The first time a Word or Excel file is opened, i.e., when either Word or
Excel have not been actively running recently (like, say for several
minutes or more), the time it takes to start the application and load
the file seem inordinately long.
Once the first
Marcello Romani wrote:
Berend Tober ha scritto:
The first time a Word or Excel file is opened, i.e., when either Word
or Excel have not been actively running recently (like, say for
several minutes or more), the time it takes to start the application
and load the file seem inordinately long.
Please review the Samba HOWTO, chapter 10, Common Errors where it
discusses this issue.
http://us4.samba.org/samba/docs/man/Samba-HOWTO-Collection/NetworkBrowsing.html#id350945
Jonathan Johnson
Sutinen Consulting, Inc.
www.sutinen.com
Aaron Kincer wrote:
Also, as others have mentioned,
I fixed this by using the following in login scripts for every share
each domain account was supposed to have:
net use driveletter: /delete
net use driverletter: \\newserver\newsharelocation
This fixed 99% of the invalid shares we had. There were a few people
that had manually mapped drives.
Berend,
This wasn't a migration from an NT domain was it? We had the problem
after a migration that starting up Office programs was incredibly slow -
it turned out there were a load of Office registry entries pointing to
UNC paths on the old Windows PDC.
Just an idea...
Cheers
Alex
--
To
My first question would be does this happen with other applications or
strictly Office? Do you get the same behavior if you attempt to open a
.doc file with Open Office?
Second, have you watched your samba logs in real time (example: tail -f
/var/log/samba/your_pc_log) as you try to open a
Alex Crow wrote:
This wasn't a migration from an NT domain was it? We had the problem
after a migration that starting up Office programs was incredibly slow -
it turned out there were a load of Office registry entries pointing to
UNC paths on the old Windows PDC.
No. Not NT. Previous file
Aaron Kincer wrote:
My first question would be does this happen with other applications or
strictly Office?
No one has complained about other apps, which in our case the next most
heavily used is AutoCAD. I think I would have heard by now if that were
a problem.
Do you get the same
Berend Tober wrote:
Aaron Kincer wrote:
...
Do you get the same behavior if you attempt to open a .doc file
with Open Office?
The answer is yes to that, but I would estimate that it is a little
less noticeable.
Let me correct that. It is a lot less noticable. Maybe even it doesn't
On Wed, 2007-06-13 at 13:03 -0400, Berend Tober wrote:
Alex Crow wrote:
This wasn't a migration from an NT domain was it? We had the problem
after a migration that starting up Office programs was incredibly slow -
it turned out there were a load of Office registry entries pointing to
UNC
Is there any difference in behavior when you open the files from within
the applications themselves so that you can remove the application load
time from the equation?
Just a little FYI--MS Office is a strange animal in how it and Samba
play together. It is not uncommon to see behavior there
How do I get this network trace that Jeremy is asking about?
James Dinkel
-Original Message-
From: Jeremy Allison
On Fri, Nov 17, 2006 at 07:50:08AM -0600, James A. Dinkel wrote:
Our samba server authenticates to Windows 2000 Active Directory and I
have ea support enabled on the share
Now copying and accessing files is plenty fast, but when setting up
permissions on directories that contain 100 GB or so of files and
subdirectories takes like 60 minutes from the time I hit Ok to the
time the permission are applied and the box goes away. This is being
set from a Windows client
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi Again,
I have traced some more on the problem.
It is the failing name resolution via netbios that delay the output from
wbinfo -u.
I can see from a trace that failing lookup's are on other DC's in the
domain, which i don't have access to, but
56 matches
Mail list logo
