[Samba] Re: Ubuntu samba slower than red hat??

[Gary Dale]

Douglas D Germann Sr wrote:
>>>       
> Gary and Guille--
>
> I checked the logs at /var/log/samba/smbd.log and have in there 
> a long series of these messages:
>
> ====clip
>
> [2006/07/25 17:36:14, 0] lib/util_sock.c:get_peer_addr(1225)
>   getpeername failed. Error was Transport endpoint is not connected
> [2006/07/25 17:36:14, 0] lib/util_sock.c:get_peer_addr(1225)
>   getpeername failed. Error was Transport endpoint is not connected
> [2006/07/25 19:12:17, 0] lib/util_sock.c:get_peer_addr(1225)
>   getpeername failed. Error was Transport endpoint is not connected
> [2006/07/25 19:12:17, 0] lib/util_sock.c:get_peer_addr(1225)
>   getpeername failed. Error was Transport endpoint is not connected
> ====end clip
>
> Does that provide a clue?
>   
Yes, but I don't know quite what it is saying. :)  It may be that it is
looking for some NetBIOS information but don't quote me on that. The
experts on this forum should be able to provide a better answer. Or you
can try googling the error line. That frequently turns up useful
information!

> I don't know anything about nsswitch nor winbind, so I suspect I am 
> not using them.
>
> Yes, I am running a firewall on ubuntu, firestarter. It has not 
> reported any incidents or events, so I suspect that is not the 
> problem. Besides it was not giving me problems (that I 
> recognized as problems, anyway) when the redhat machine was the server.
>
> To answer Gary's question: This is a small office involving 3 people 
> and 3 Ubuntu clients, one WinXP Pro client, and 2 Win95 clients. 
> This is a production environment using mainly word processing and 
> spreadsheets. The linux boxes connect using fstab entries like this:
> ===clip
> //earth/vol1        /sam/vol12  cifs   
> rw,user,credentials=/root/.smbcredentials,uid=doug,gid=apps       0       0
> ===end clip
>   
I suggest you set up your Samba server as a domain controller. This
means that it is a central login point that keeps track of the user
accounts. SWAT has a wizard to do the basic setup for this.



> Memory on the old eMachines is 128Mb; on the new machine it is 256Mb.
>   
This should be enough for running a server without a GUI. However, you
could bump it up to as much as it can hold and you should see a
performance increase. Even slow memory is faster than a disk drive, so
the more you can keep cached, the better.

> When you say set the log level up higher, the only thing I see is 
> to log more to syslog instead of to the samba logs, if I am reading 
> it correctly. It is set to 0 right now. Is this what you mean? 
> What should I try setting it to?
>   
10 or 100 should get you lots of messages.

> The files were moved to the new server by way of tgz files. These 
> were created by a backup Ubuntu machine. The data HDD on the old 
> server crashed, necessitating the use of another machine.
>   
This means the old user and group ids are being used. In Linux/Unix,
there isn't a lot of connection between the user number and the account
name (same with groups). User 1001 is whatever account name is in
/etc/passwd at the time. You're going to have to either change the
account names and groups to match your old setup or change the
owner:group settings to match your current accounts and groups.

> When I run top, I presume I run it on the server. What sorts of 
> things am I looking for? I am getting things like this:
>
> ===clip
> Tasks:  72 total,   2 running,  70 sleeping,   0 stopped,   0 zombie
> Cpu(s):  0.3% us,  0.0% sy,  0.0% ni, 99.7% id,  0.0% wa,  0.0% hi,  0.0% si
> Mem:    507432k total,   503136k used,     4296k free,     3984k buffers
> Swap:  1485972k total,    18932k used,  1467040k free,   301640k cached
>
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
> 10124 doug      16   0  2200 1088  856 R  0.3  0.2   0:00.18 top
>     1 root      16   0  1568  532  460 S  0.0  0.1   0:01.09 init
>     2 root      34  19     0    0    0 S  0.0  0.0   0:00.00 ksoftirqd/0
>     3 root      RT   0     0    0    0 S  0.0  0.0   0:00.00 watchdog/0
>     4 root      10  -5     0    0    0 S  0.0  0.0   0:04.76 events/0
>     5 root      10  -5     0    0    0 S  0.0  0.0   0:00.01 khelper
>     6 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kthread
>     8 root      10  -5     0    0    0 S  0.0  0.0   0:01.21 kblockd/0
>     9 root      20  -5     0    0    0 S  0.0  0.0   0:00.00 kacpid
>   147 root      15   0     0    0    0 S  0.0  0.0   0:00.08 pdflush
>   148 root      15   0     0    0    0 S  0.0  0.0   0:00.02 pdflush
>   150 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 aio/0
>   149 root      15   0     0    0    0 S  0.0  0.0   0:27.80 kswapd0
>   751 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kseriod
>  1798 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 ata/0
>  1799 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 ata_hotplug/0
>  1802 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 scsi_eh_0
> ===end clip
>
> I just saved a document and there was briefly a line for smbd.
>   
This shows that your server is operating normally. Nothing is chewing up
CPU cycles.

>   
>> Is this a domain controller? Does it farm out password checking to 
>> another server? You may want to set up SWAT and use the wizard to set up 
>> the server in its intended role (domain controller, member server or 
>> stand-alone).
>>     
>
> I do not know enough to answer these questions. I used webmin and swat 
> to tweak the old server, and have installed swat on the new but not 
> yet run it.
>
> Thanks folks! I feel like you are getting me on solid ground.
>
> :- Doug.
>
>