Re: [Samba] Questions about PDC with SAMBA
Marcelo Opazo Vivallos: Hi! I have 8 subnets: 192.168.100.x/24 192.168.150.y/24 192.168.200.z/16 etc ... Install a Primary Domain Controller (PDC), so that users to perform authentication on the domain with its mounting remote disks, among others. The network is correctly configured, that is, the teams are perfectly by ping and the open ports between the networks see no problems. Notes: - I disabled the multicast UDP in the route - Add the lines in the smb.conf: - Hosts allow = 192.168.100. 192.168.150. 192.168.200. [..] The others go right, because as I said it all works this, but only for the subnet. - Customers and resolve the server name netbios (wins OK) However, I encountered only heard the PDC requests only on its subnet (192.168.200.z). * Questions: - Why if the machines are visible between them they are not able to hear the requests to the PDC? - Is there another method instead of putting 8 samba servers in each subnet? Reggard Marcelo Opazo Vivallos Chile Hi. Try to use LMHOSTS file on Windows clients. We had problems when clients in some routed subnets can not reach the PDC sometimes even though the WINS was configured OK - and this helped us. We are using this batch file on Windows: echo 192.168.x.y PDC_NAME #PRE #DOM:DOMAIN %systemroot%\system32\drivers\etc\lmhosts echo 192.168.x.y DOMAIN \0x1b #PRE %systemroot%\system32\drivers\etc\lmhosts nbtstat -R Replace PDC_NAME with netbions name of your PDC and DOMAIN with your domain name and 192.168.x.y with IP of the PDC. Length of record on the second line must be kept at 16 characters. Vlastimil Setka -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/options/samba
Re: [Samba] smb keeps stale connections from logged out users and shut down clients
Volker Lendecke wrote: On Mon, Feb 09, 2009 at 03:17:27PM +0100, Frederik wrote: We are using Samba 3.0.33 as a PDC and file server for Windows XP clients, but seem to be suffering lots of stale connections. smbstatus still shows connections from users already logged out and even from systems which have already been shut down. Restarting the samba service, cleans up all stale connections. We tried deadtime = 5 in the samba configuration file, but this does not help at all. The parameter keepalive doesn't help either? Volker We have similar problems with Samba 3.0.24 (Debian Etch) as PDC for ~120 Win XP clients. See also this thread refering about our problems: http://lists.samba.org/archive/samba/2008-September/143701.html Since that time I made some investigations with interesting results. About once per month it's a locking issue and the smbd proccess exists. It is not a big problem. After Samba restart all is OK. But other problem come up when logging smbstatus output every 10 seconds for audit purpose. Wery often smbstatus shows a connected user from computer which have already been shut down or from computer where other user is loged in (see referred thread). This connections is showed usually for ~30 seconds and then disappear. I modified my logging script: it also logs the proccess name of each PID in smbstatus. I found that PIDs of the ghost connections is not the smbd! It's cron tasks, apache cgi scrips, ... ! I patched smbstatus: normally it loads the sessionid.tdb database and checks if PID of proccess in database exists. In my modification it also shows records with PIDs that do not exist. After a day after the Samba restart and ~400 domain logon/logoff there is about ~15 records with bad PID in the sessionid.tdb - with normal smbstatus this records are not showed - only if the process with the PID coincidentally exists... I think that Samba should have some self-healing mechanisms with sessionid.tdb. I can imagine some garbage collector periodically checking existence of PIDs for all records in sessionid.tdb and if such PID does not exists it should delete the record. Have Samba such feature in newer versions, or is it other error? Thanks. Vlastimil Setka -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/options/samba
Re: [Samba] samba server in two lans
Mohammad Reza Hosseini napsal(a): hello we have a samba server on centos 5.2 and 2 different lans. so we gave the server to ips eth0:172.16.93.217 and eth1: 192.168.89.3 but after this when we tried to join clients (windows xp) to the domain the error: the specified domain either does not exist or could not be contacted. what is the solution? I had the same problem with Samba 3.0.24 - Debian Etch package. On PDC server with 5 interfaces (VLAN) when I tried to join clients to domain, sometimes I got several strange errors. Sometimes that errors came on at logon... From tcpdump output I found a problem that Samba server sometimes send browse-reply UDP packets with source IP address of other interface than the outgoing interface. So the client can't locate PDC address. This solution perfectly works for me: smb.conf - global section: interfaces = 192.168.1.0/24 lo socket address = 192.168.1.5 where the 192.168.1.5/24 is address of one local interface. Be ware that now Samba can be reached only on this 1 address. On the clients is required to set the LMHOSTS file, so client knows selected IP of PDC. I'm using this batch: echo 192.168.1.5 PDCNAME #PRE #DOM:DOMNAME %systemroot%\system32\drivers\etc\lmhosts REM keep length = 16 chars including the \0x1b echo 192.168.1.5 DOMNAME\0x1b #PRE %systemroot%\system32\drivers\etc\lmhosts REM reload config nbtstat -R Maybe this issue is solved in some newer Samba version. -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Closing sessions and smbstatus
Hi, When are client sessions closed? Let me explain what I'm trying to do... we're in a School district and we try to stop kids logging more than once. They way I did this before was to dump the active sessions from our previous Server2003 fileserver into a file once a minute and process it with a Perl script to check who was connected from where, rebooting machines remotely as needed! This work well enough with the odd 'hung' session causing minor problems. I'm in exactly the same situation. The school, PDC for ~100 computers, hundreds of users. We need to track the logon / logoff. I can't find any usable tools so I made my own system. I found that most reliable is the smbstatus output. Windows do strange thinks with connections during domain logons so use of preexec script is complicated. By the Perl script I run smbstatus every 5 seconds, scan the changes from previous run and write it to the MySQL DB. That's all woks fine. So now I'm trying to do the same thing with our new Samba (3.0.31) fileserver using the output from smbstatus. However, in many cases sessions are still in there long after the user has logged out of the machine. I'm even seeing two sessions for different people on the same machine with the same pid number! How is this working? Why are not all sessions ending when the user logs off? Am I going to be able to use this for what I'm trying to do?!! I have some problems with this too. See this thread: [Samba] smbstatus - switched off computers are sometimes showed http://lists.samba.org/archive/samba/2008-September/143701.html Now I get some new experience with it. The main problem is that samba sometimes doesn't update the sessionid.tdb file when the process exits. This records is not showed in smbstatus output, because smbstatus checks if the PID exists. I patched the smbstatus so it showed me that there is the records with no related PID. Then, maybe after 1 day or so, this PID is used for other proccess and I can see the ghost logon in my tracking system (and in most cases logoff at next run - after 5 seconds). On the list is now the thread [Samba] processes not closing where is described some self-repair function related do sessionid.tdf file. The samba process when writing to this file should check all records and delete it if the PID doesn't exist. It will be nice but In my situation it doesn't work. Maybe it's because of Samba version (3.0.24, official Debian Etch package). The most strange think I've seen is that I get some fake logon records for one user day-by-day at the same time. Let say [EMAIL PROTECTED] - tracked logon at tuesday 14:10:12, then at the same time at wednesday and thursday. In fact the COMP1 is switched off or other user is loged on at the time. The USER1 were loged on the COMP1 at monday. The fileserver itself is working great, we have over 2000 users happily using it with less problems than we had on the Windows box. I really appreciate all the work the Samba team does! The same experience. Samba-based solution with one PDC is rock-stable for us in comparsion with several Windows 2003 AD servers running before. Many thanks, Steve Rippl Woodland School District -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Closing sessions and smbstatus
But checking our server I find that the processes DO still exist! So I'm getting a user session in smbstatus with a specific PID and when I ps -ef | grep PID there is the smbd process still running, yet the user has long since logged out (days ago)?! It's not just that the .tbd file hasn't been updated, it's that the /usr/local/bin/smbd process is still running. A restart of Samba clears them all up immediately, but why are they hanging around when the client is gone? Is this just happening to me on our particular setup or is this normal behavior? We have ~400 logons per day. Sometimes (average 1 process per day) some processes hangs - the PID exists but user is several hours loged off. In this time other users were loged on this station... But there are some locked files connected with this PID - I think this is because the process isn'n closed. Can you see some locked files connected with the bogus PID in smbstatus output? Thanks. -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] smbstatus - switched off computers are sometimes showed
Volker Lendecke wrote: On Fri, Sep 19, 2008 at 10:02:02PM +0200, Vlastimil Šetka wrote: I have a Samba server configured as PDC for ~100 computers. It's version 3.0.24 running on Debian Etch (distribution package). I want to write a tool for user logon/logoff tracking. Because parsing log files is hard job (windows frequently disconnets or connects during user session or etc.) I decide to use smbstatus output which seem to be reliable. So I run smbstatus binary every 10 seconds, parse the output and all is fine. But sometimes in the smbstatus output is show computer which is several hours switched off. At the next run (after 10s) this record is not presented. Sometimes this ghost record is shown for ~30 seconds. Does that smbd still exist? If computers are just switched off then Samba might not notice. I plan to check it but it's complicated because the ghost record is presented only for several seconds. I have to log the PID and related ps output. Smbstatus checks if the proccess with the PID still exists. So if it's showed the smbd probably exists. The computer is really switched off for several hours when this issue is presented. I also check the log file and there is no activity at this time. Today there were two ghost records for two different users on one computer - this computer is switched off for 24 hours and this two users were loged in during last 3 days. Example: 11:10:21 - MACHINE1/USER1 - is in smbstatus ... constantly is in smbstatus 12:14:11 - MACHINE1/USER1 - is NOT in smbstatus 12:15:21 - MACHINE1/USER2 - is in smbstatus ... 12:45:31 - MACHINE1/USER2 - is NOT in smbstatus Now the MACHINE1 is switched off. 18:16:21 - MACHINE1/USER1 - is in smbstatus !! (and the MACHINE1 is switched off several hours!) 18:16:31 - MACHINE1/USER1 - is NOT in smbstatus Sometimes (cca twice a day) smbstatus return error output like this: tdb(/var/run/samba/sessionid.tdb): rec_read bad magic 0xd9fee666 at offset=116988 This is normal. Smbstatus reads the tdb files in read-only mode. This implies no locking which in turn might lead to smbstatus reading an inconsistent tdb file. Yes. In sources I notice that no locking is used. Smbstatus only reads the file so it's OK. I also want to use INOTIFY on sessionid.tdb file but it's not usable because smbd has the file constantly open for write. Volker -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba
Re: [Samba] Re: smbstatus - switched off computers are sometimes showed
Avery Payne wrote: Vlastimil Šetka wrote: Sometimes (cca twice a day) smbstatus return error output like this: tdb(/var/run/samba/sessionid.tdb): rec_read bad magic 0xd9fee666 at offset=116988 It looks like you have an issue with a record. Stop your service for a moment, go in and cp your files in /var/run/samba to a new directory, say to /var/run/samba.backup, then go into /var/run/samba and do this: tdbbackup -sbak *.tdb tdbbackup -v -sbak ...then start Samba. The first makes a backup of your tdb tables; the second verifies (and restores if needed) your tdb records. This is safe, but if there was any issue, you can always restore your files from the copy in /var/run/samba.backup that you made. At first: I get this error message only about twice a day (smbstatus is executed every 10 seconds) and the offset number every time changes. But the issue with a ghost user is more frequent. I read the source of smbstatus and I think that delete /var/run/samba/sessionid.tdb will be enough. In my tool I use smbstatus -p and in this case any other file is not used. I already tried to stop samba, delete this file and then start samba. New empty sessionid.tdb was created but nothing changes. Maybe I will try your advice about delete and backup all tdb files but I have to wait for some time for testing when the system is not in use. It's a bug or a feature and smbstatus is only informative? It's solved in some newer version? Can you tell me about better source for connected user status? On RHEL52 here and just moved up to 3.0.28 as part of an update from stock RHEL50. Significant difference in stability and behavior from the stock 3.0.25b. Twice-a-week locking issues have just disappeared and using the Computer Management tool attached to the Samba server shows actual files open instead of ghost files. Does apt-get update apt-get upgrade show any entries for Samba? The package is up-to-date Etch version. -- To unsubscribe from this list go to the following URL and read the instructions: https://lists.samba.org/mailman/listinfo/samba