headers
Jamrock | 1 May 2005 02:52
Picon
Favicon

LDAP shared files error

I am installing Samba 3.0.14a on Whitebox Linux 3.0 with Openldap 2.2.24.

Openldap is working well and we can query the entries that we have in our
addressbook.

I have installed Samba from source.  The configure, make and make install
appear to go okay.

When I try to start Samba
with /usr/local/samba/sbin/smbd -D
I ge the following error

/usr/local/samba/sbin/smbd: error while loading shared libraries:
libldap.2.2.so.7: cannot open shared object file: no such file or directory

I see the libldap.2.2.so.7 in /usr/local/lib and
/usr/local/openldap-2.2.24/libraries/libldap/.libs

My include/config.h file contains the line
#define HAVE_LDAP 1

What am I missing?

--

-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba
Permalink | Reply | 2 comments |
headers
Michael Joyner | 1 May 2005 02:10

every increasing connection count followed by load # overload

I am having a bizarre problem with a samba server here at Edward Waters 
College.

Hopefully someone can give me insight to inform me what to look for.

samba version = 3.0.14a
SuSE 9.1, up-to-date-via-apt
AD/W2K3 (no sp)
Winbind using RID algorithm

server purpose: HOMES, DESKTOPS, MY DOCUMENTS, PROFILES
Quotas are: ON
Filesystem: Reiser
HW Raid 5 (MegaRaid)

4 logical processors (dual+ht).
uname -a:

Linux ewc05 2.6.5-7.151-smp #1 SMP Fri Mar 18 11:31:21 UTC 2005 i686 
i686 i386 GNU/Linux

If the number of connections exceeds 155 (per samba counting), the 
system load quick escalates to like 700+, and the user count also 
suddenly skyrockets. a majority of the samba processes seem to get stuck 
in state "D"

inode count usage is unrealisticically high. :)

I deleted all the samba db files, no change.
(Continue reading)

Permalink | Reply | 4 comments |
headers
Doug VanLeuven | 1 May 2005 02:26
Favicon

Re: Folder Redirection broken if access is from ACL only

Grant Bigham wrote:

>I have an issue with W2K/XP using Folder Redirection to a Samba homes
>share (or any share for that matter). This is only a problem when access
>for a user is via an ACE (ACL) and not the traditional file system
>permissions.
>
>The problem is on Linux (various distribs (SLES8 and FC2) 2.4 and 2.6
>Kernels), and Samba-3.0.11 on ext3 file systems mounted with
>user_xattr,acl options. 
>
>This is not an ACL problem as such. Access to shares and the data within
>is fine using ACLs, it only becomes a problem when Windows tried to
>access redirected folders on Samba, where that access is granted via
>ACLs only.
>
>So for example (user is cath in this example):
>
>[root <at> gandalf users]# ls -ld cath
>drwxrwx---+ 5 root root 4096 Apr 15 20:40 cath
>
>[root <at> gandalf users]# getfacl cath
># file: cath
># owner: root
># group: root
>user::rwx
>user:cath:rwx
>group::---
>mask::rwx
>other::---
(Continue reading)

Permalink | Reply | 0 comments |
headers
Craig White | 1 May 2005 02:32

Re: every increasing connection count followed by load # overload

On Sat, 2005-04-30 at 20:10 -0400, Michael Joyner wrote:
> I am having a bizarre problem with a samba server here at Edward Waters 
> College.
> 
> Hopefully someone can give me insight to inform me what to look for.
> 
> samba version = 3.0.14a
> SuSE 9.1, up-to-date-via-apt
> AD/W2K3 (no sp)
> Winbind using RID algorithm
> 
> server purpose: HOMES, DESKTOPS, MY DOCUMENTS, PROFILES
> Quotas are: ON
> Filesystem: Reiser
> HW Raid 5 (MegaRaid)
> 
> 4 logical processors (dual+ht).
> uname -a:
> 
> Linux ewc05 2.6.5-7.151-smp #1 SMP Fri Mar 18 11:31:21 UTC 2005 i686 
> i686 i386 GNU/Linux
> 
> If the number of connections exceeds 155 (per samba counting), the 
> system load quick escalates to like 700+, and the user count also 
> suddenly skyrockets. a majority of the samba processes seem to get stuck 
> in state "D"
> 
> inode count usage is unrealisticically high. :)
> 
> I deleted all the samba db files, no change.
(Continue reading)

Permalink | Reply | 1 comment |
headers
Michael Joyner | 1 May 2005 02:55

Re: every increasing connection count followed by load # overload


/mjoyner <at> vbservices.net
> ----
> don't know the answer but just an observation - that is a pretty old
> kernel (and early 2.6 version) - doesn't SuSE have something in the
> 2.6.10 variety?
> 
> Craig
> 

Hrmmmmmmmm, let me check my apt setup. :)
--

-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba
Permalink | Reply | 0 comments |
headers
Detlef Grittner | 1 May 2005 03:46
Picon
Favicon

browsing list of nmb not correctly updated

I have set up a network of two Linux machines, which communicate with
Samba and nmb, as if they were Windows clients in a peer to peer network
(Workgroup).
That is both machines have a Samba server and a client.
Nmb is active as well and both machines have "preferred master" set to
"auto" and "local master" to yes.

This worked with older Samba versions, but with Samba version 3.0.13 I
run into problems.
Now the master browser on the network only knows about himself.
That is only the netbios name of the local machine is in the browsing
list, if this machine is the master browser.
So I can see only one machine in my network of two Linux machines
instead of two and only the Linux machine with master browser is the
visible one.

If I plug in a third machine running on Windows XP and it becomes the
master browser, then all three machines are visible in the Workgroup.

Where shall I begin to look into the problem of my two Linux machines?

Detlef
--

-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba
Permalink | Reply | 0 comments |
headers
pb | 1 May 2005 05:10
Picon

smbclient - recently broken?

Hello all,

forgive me if this is old news but it seems to me that
smbclient is prone to getting into an infinite loop
when mget'ting a stack of files.

I have tested (compiled from source with defaults)
samba versions 3.0.10, 11, 12, 13, 14, 14a
Versions prior to 3.0.13 are fine, in my environment, versions 3.0.13 
and later are causing the problem.

The basic data on my system is
-a P4 3.2 Ghz, 512 Ram, Slakware 10, Intel Etherexpress Pro Nic. 	this 
server is very lightly loaded.	
-The client is a NT4 with SP6a, with a 100 day uptime habit 			(i.e. 
really reliable). It is the current file server and is
backed up daily with smbclient.

I had client problems with Win98 clients with samba v3.0.10,11 and 12 
during this testing, but I have a stable situation with samba v3.0.7 
where both smbclient as well as the Win98/XP Pro clients work fine.

I can provide packet captures of the infinite loop if needed.

Peter.

--

-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba
(Continue reading)

Permalink | Reply | 4 comments |
headers
Ephi Dror | 1 May 2005 05:15

wbinfo -n and wbinfo -s are broken with win2003 SP1

Hi All,

I upgraded my windows 2003 enterprise server  to sp1 and having problems with winbindd/wbinfo.

I am using samba 3.0.6 and joined as ADS

wbinfo -u 
wbinfo -g
wbinfo -p
wbinfo -t 

work like a Swiss watch

However 

wbinfo -n ephi

wbinfo -s .....

DO NOT WORK.

Is the latest SAMBA release (3.0.15pre2 and up) cover those issues (-n and -s) 

If we joined as DOMAIN (NT4 style) even other flags such as -u, -g didn't work well.

Thanks so much for all your help. We relying heavily on those 2 wbinfo options.

It sounds scary to me. Can we get updates on the situation.

Cheers,
(Continue reading)

Permalink | Reply | 0 comments |
headers
Michael Lueck | 1 May 2005 01:37
Favicon
Gravatar

Re: Q for folks using Samba with XFS

Greg Freemyer wrote:

> xfsdump works on an entire filesystem/partition.  I don't think is
> allows dumping of a directory tree.

Bing bing bing!!! You got the prize Greg, thanks!

-- 
Michael Lueck
Lueck Data Systems
http://www.lueckdatasystems.com/

Remove the upper case letters NOSPAM to contact me directly.

--

-- 
To unsubscribe from this list go to the following URL and read the
instructions:  https://lists.samba.org/mailman/listinfo/samba
Permalink | Reply | 0 comments |
headers
Eric Feldhusen | 1 May 2005 06:56

Re: samba 3 performance

John H Terpstra wrote:
> To add to this info. The last benchmarks I did were in 2003. Within the next 
> two months I will benchmark a new system that will have dual 3Ware SATA RAID 
> controllers each with 6 high performance drives in an Opteron system. I am 
> anxious to see the performance stats, particularly compared against the 
> previous stats on an AMD dual CPU system with a single 3Ware IDE RAID 
> controller and 4x60GB WD 7200 rpm drives - 452MBytes/sec peak I/O with samba, 
> and a peak sustainable write rate of 115 MBytes/sec. That write rate nose 
> dives badly with concurrent mutiple file write activity and/or read activity 
> that causes significant seek activity on the drives in the RAID array.

I have two servers running Samba 3.0.4 on Redhat 8.0 on dual athlon 
1600mp on tyan s2469 boards each using 3Ware 7504, with 3 200GB 7200rpm 
WD in a hw raid 5 array using ext3.  Now, while I don't have to all the 
numbers to really quantify what happened on my network, I saw similar 
problems with my setup with concurrent read/write activity that causes 
high seek activity.  On my networks, I originally had all windows 9x 
machines, then, at both schools we swapped out about 1/2 of about 200 
machines with windows 2k machines,  suddenly at beginning of class 
periods and end of class periods, high loads, 40-50 via looking at top, 
on the server, but very little cpu and network use, but high disk 
activity ( http://webminstats.sourceforge.net ).  Anyway, in looking at 
3ware's site, I found a reference that with ext3, the 3ware card is 
slower than XFS or ReiserFS ( 
http://www.3ware.com/KB/article.aspx?id=10095 ). Now, unfortunately, 
it's not an apples to apples comparsion, but I just finished wiping out 
and upgrading both servers to FC3 on XFS, on the exact same hardware, 
and I'm seeing a much lower load (3-7 via top) at the same class changes 
and the system remains responsive.  I wanted to wipe the system and 
reinstall with RH8 with XFS with same hardware, but with production
(Continue reading)

Permalink | Reply | 0 comments |

Gmane