cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Announcements

Welcome to the Cisco Small Business Community

Have a question? Click on a topic board below to get started in the community.
Get the latest news in this issue of the Cisco Small Business Monthly Newsletter

2294
Views
0
Helpful
10
Replies
Highlighted

NSS6000 drops users - samba (No space left on the device)

I have a NSS6000 joined to a Windows 2003 Domain with 25 users. The users  (windows xp) are connected to the NSS6000 directly with a mapped drive.  Sometimes NSS6000 drops off users and they are not able to store opened files or  reconnect to the mapped drivers.

The share volume is named general_shared with 800GB and near of 19% is free  space. The firmware is 1.20 version. The time is sinchronized with the DOMAIN  server with NTP and updates correctly. When i look for something strange in the  logs only i can find the next message infinitely repeated:

Mar 16 18:05:10 NSS6000 smbd[7897]: [2011/03/16 18:05:10, 0]  lib/util_tdb.c:tdb_log(664)
Mar 16 18:05:10 NSS6000 smbd[7897]:    tdb(/var/lock/samba/connections.tdb): expand_file write of 1024 bytes failed (No  space left on device)
Mar 16 18:05:10 NSS6000 smbd[7897]: [2011/03/16  18:05:10, 0] smbd/connection.c:claim_connection(208)
Mar 16 18:05:10 NSS6000  smbd[7897]:   claim_connection: tdb_store failed with error Record does not  exist.
Mar 16 18:05:10 NSS6000 smbd[7897]: [2011/03/16 18:05:10, 1]  smbd/service.c:make_connection_snum(897)
Mar 16 18:05:10 NSS6000  smbd[7897]:   too many connections - rejected


Then I have to restart the NSS6000 and it returns to work correctly and  all users are able to access. I attach a bigger slice of the samba.log file  showing this situation (a correct access from network scanner, a user that tries  to access, some restarts of samba service and the message repeated). All other  logs doesn't show any error.

Does anyone find with this situation? What can i do to correct this?

Regards,

Javi.

10 REPLIES 10
Highlighted
Frequent Contributor

Hi

Javi.

Just from the logs it looks like the max amount of connections are being reached ? how;with only 25 connections is the question since by default the NSS6000 supports over 100 CIFS. Under the SHARE tab Property tab is your NSS set to (STANDARD or ADVANCED) also currently running firmware ?.

Thanks,

Jason Bryant

Cisco Support Engineer

.:|:.:|:.

Highlighted

Hi Jason,

Normally 25 pc's are connected with the NSS6000 but each user is able to open more than one file concurrently. Is limited the maximum number of concurrently opened files? What is the maximum value?

The connection profile is set to "Advanced". At past, when the connection profile was set to "Standard" this problem repeats more frequently. I changed this as suggested in this forum. Other default configuration changed where user/group settings. I set the values that are show in the attached file.

Thanks,

Javi.

Highlighted

Hi javi,

Can you tell me the setting of your CIFS idle disconnect time is ?

This is located under Shares--> CIFS setup

Thanks,

Jason Bryant

Cisco Support Engineer

Highlighted

Hi Jason,

Idle disconnect time is set to never. The values of this attribute goes from 2 to 30 minutes. If i set any of these values, when idle disconnect time past, users mapped drives become disconnected and fails any shortcut to a file in the storage.

Thanks,

Javi.

Highlighted

Javi,

This could be your problem since idle time is set to 0 then every connection that is made to NSS is kept alive. Their are only so many connection that can be made to nss, once this is reached then you will get your error. I will check on the disconnection of MAPPED drives, the user shouldn't lose connection to that mapped share.

Thanks,

Jason Bryant

Cisco Support Engineer.

Highlighted

Javi,

Please enable idle disconnect time on the shares to at least 5 minutes. This should solve your problem with the errors; now with your mapped drives disconnecting. I have tested and even way after my idle disconnect time and my mapped drive never disconnects. I think it might be a problem in how you mapped your drives.

Topology view

Can you show me how you mapped your drives and which settings you use and file path.

Also are you using local users or a domain?

Thanks,

Jason Bryant

Cisco Support Engineer

Highlighted

Hi Jason,

I'm sorry by the delay in the response but i wanted to change the configuration  of the idle disconnect time, perform the firmware update and take some days of working users experience.

At first i'm going to explain shortly the topology of the network:

- A Microsoft Domain Server controls all users access. This domain server performs the roles of dns and dhcp. Users are logged to this server. In the netlogon script i've configured the next sentences to create mapped drives.

net use k: \\SERVER\Users\%USERNAME%
net use l: \\NSS6000\shared
net use m: \\NSS6000\shared2

- A NSS6000 network storage server that is joined to the domain. Users acces thought network drives. Users, server and storage are in the same network with a GBEth Switch.

- A NSS4000 network storage server to store backup's.

When i performed the change of idle disconnect time and firmware update all works so fine with the only problem that, when idle time expires, network drives were showed as disconnected. I attach a screenshoot of this case in the server (in user's computers are the same). Despite this, users double click over mapped drives and the connection is restored and data is accesible. Excepts  the 2 May that i'd to reboot the NSS6000 the rest of days works so fine.

This morning i ran a backup copy of the contents of the NSS6000 to the NSS4000 (incremental copy). The backup is performed with an application in the server.  This application in the server copy data from NSS6000 to the NSS4000.

This application, at first, navigates file tree to find files that are added in the incremental copy and then, performs the copy. Past aprox. 50% of files then  NSS6000 starts to drops all users connections and the copy starts to show access errors. Then, the only solution is to restart the NSS6000.

At next is show a part of the CIFS log. Continuously the next lines are show:

May 11 11:46:25 NSS6000 smbd[30307]:   SERVER (10.0.0.15) connect to service shared2 initially as user DOMAIN\User (uid=40500, gid=40513) (pid 30307)
May 11 11:46:25 NSS6000 smbd[30307]: [2011/05/11 11:46:25, 0] lib/util_tdb.c:tdb_log(664)
May 11 11:46:25 NSS6000 smbd[30307]:   tdb(/var/lock/samba/locking.tdb): expand_file write of 1024 bytes failed (No space left on device)
May 11 11:46:25 NSS6000 smbd[30307]: [2011/05/11 11:46:25, 0] lib/util.c:smb_panic(1633)
May 11 11:46:25 NSS6000 smbd[30307]:   PANIC (pid 30307): Could not store share mode entry
May 11 11:46:25 NSS6000 smbd[30307]:   
May 11 11:46:25 NSS6000 smbd[30307]: [2011/05/11 11:46:25, 0] lib/util.c:log_stack_trace(1737)
May 11 11:46:25 NSS6000 smbd[30307]:   BACKTRACE: 1 stack frames:
May 11 11:46:25 NSS6000 smbd[30307]:    #0 [(nil)]
May 11 11:46:25 NSS6000 smbd[30307]: [2011/05/11 11:46:25, 0] lib/fault.c:dump_core(181)
May 11 11:46:25 NSS6000 smbd[30307]:   dumping core in /var/log/samba/cores/smbd
May 11 11:46:25 NSS6000 smbd[30307]:

The server is the only machine connected to the NSS4000. The NSS4000 unit doesn't disconnect users (idle time is set to never) and the CIFS log doesn't show the NSS6000 situation and mapped drives never are show as disconnect (N, O and X drives of the screenshoot).

Do you have any suggestion? I'll try to make a quickly response.

A lot of thanks,

Javi.

Highlighted

Javi,

Quick question how is the NSS6000 connected to the network? Are you using both nic in bond/(802.3ad mode) (active backup) or standalone?

Also how is the NSS4000 connect also?

Thanks,

Jason Bryant

Cisco Support Engineer

Highlighted

Jason,

NSS6000 and NSS4000 are both configured with Active Backup (failover) bounding mode with and MTU of 1500 bytes with the 2 ports connected to 2 different switches at GbEth.

Thanks,

Javi.

Highlighted

Disconnect time was changed to 5 minutes with user's disagreement (because shared units become disconnected). At last week, the "PANIC error" was repeated every day.

Any suggestion?

Thanks!!!