red

I used midnight commander to copy the file, not a command as you showed. RED

Success!!! It was not my config files at all, they were fine. It turned out to be their location and permissions. I installed BIND from the RPM "bind-9.3.1-4.320060mdk" and "bind-utils9.3.1-4.320060mdk" as provided onteh CD or through the RPMdrake. This installlation was putting the named.conf and resulting <zone> file in /var/lib/named/etc. To add to the problem, Webmin which I was using to configurre the named.comf and zone files was editing the files in /var/lib/named/etc After looking at the messages log and seeing that named was not finding the named.config file when starting, just for th ehell of it i copied the named.conf from /var/lib/named/etc to /etc/ and the error message changed from file not found to permission denied. some additional searching on this message indicated that permissions for the folder containing named.conf had to be modified so that the named group had full access. This was necessary because even trying to launch named as root failed because of permissions, which was supported by a post found on another site that indicated tha named trumps roots ability to override some file permissions and the named group mus be specifically granted permission. This is where I think that while this is working, it is incorrect. I had to change permissions on the /etc/ folder. I changed the group from root to group and gave it read and write access, when I did this, named fired right up and off we went. The <zone> file even updated from my windows DNS server!!!. This is one of my pet peeves with using linux, while some things are very consistent, others are not and this sort of issue should not have happened. the named script should be looking in the correct locations for the config files and I cetainly shouldnot have had to change permissions on the /etc/ folder. Any idea where to change what so that I can use the files in /var/lib/named/etc as installed instead of /etc/ and restore permissions? Thanks

Well I did end up with SOME success. I found out that the config files and data files were not in the places the named script expected them to be, I moved them and named finally actually ran. I was able to get the firewall to forward queries it could not answer to the internet, I just could not get it configured to answer for my local domain (domain.local) So now its a matter of configuring this zone properly. My set up is like this: Domain.local and domain controller (windows primary DNS)-->firewall/router (Mandriva 2006 box w 2 NICS)--> Internet I want to make the firewall/router a DNS as well to answer internal queries and forward internet queries. Thanks for the help.

the firewall is a Mandriva 2006 as stated in the original post. I managed to get some response yesterday after massing with it some more, but I am unsure if it was a fluke or something I actually figured out. The config files don't seem to be in the right places, so the more information you can give will be appreciated. Thanks.

Hiya folks, I am having one heck of a time getting this to work. I have a Mandriva 2006 Free system(firewall) set up as a firewall, and working just fine. (2 NICS, one internal, 1 external) Behind this firewall I have a Windows Active Directory domain which has the DC set up as a DHCP server and DNS server. Right now, I have the DNS on the DC answering queries for the local domain and forwarding queries it cannot answer (internet) to my ISPs DNS and everything is working perfectly. My only problem is that all users lose the ability to reach sites (internal and internet) when I take the DC off line for maintenance because the DC is the only DNS. I want to set the firewall up as a secondary DNS to answer queries (internal) just as the Windows DC doesn and forward queries it cannot answer (internet) to my ISPs DNS. I have tried 7 ways to Sunday to get this working and I just cannot seem to get BIND to answer anything. I have found a numbe rof how-tos but none seem to have the right answer, or I just dont get the config right. for the sake of clarity, lets use the follwoing variables so that any information you can give regarding configuration is understood: Windows DC Server = WDC WDC Name = SERVER1 Windows Domain Name = domain.local Mandriva 2006 Box - firewall ISP domain = ISPdomain I woudl also need to know which ports to open in the firewall, as I have it pretty locked down and suspect that I will need to open things up a bit for DNS and resulting traffic to occur between firewall and domain.local. Thanks, RED

Thanks for all the replies. This thread can be considered solved or closed. RED

I have a 2005 LE box set up as a firewall/router currently with a P2 450 on an intel based MB. I acquired an Athlon 750 with VIA chipset and wondered if I will run into any problems just swapping the hardware out. Is there a different kernel for the Athlon? Will the os configure itself for the new chipset etc? Will a reinstall resolve any issues I encounter or do I need to start from scratch? Thanks in advance for the input. RED

Becasue of the current state of updates (bein unavailable for 9.2) I have been using 2005 LE with Virtual PC and am ready to upgrade my 9.2 box. My question is, will this upgrade even work or will I experience problems. Should I just whack the drive and do a clean install? Thanks, RED

Just wondering how to view the NDS information provided by my ISP's DHCP server. I have been expereincing some appearant DNS issues lately and need to know the IP addresses of the DNS servers I am being provided my my ISP for my internal DNS forwarder. Thanks in advance. RED

ya know, it still sounds as if tapes are the way to go with this,I just wish I could have convinced him to fork over the dough for an adequate tape drive. Though I could limit the size of the shared folder ultimately limiting the size of the backup and allow me to purchase a smaller tape drive. Once the share gets full, they would be forced to archive their older or unused information, which could have a huge housekeeping benefit as well. I appreciate all the input. Thanks RED

I am using Postfix...sendmail I just couldn't seem to get configured to do anything at all. This is exactly where I need help with configuration. It seems postfix is actually sending the message but my ISP mail server and web host mail server are rejecting it. Oddly enough it went through my hotmail account...go figure. I keep seeing "relay" pop up in messages and web searches relating to this issue and I am unsure that is my desired behavior. Maybe i should explain what i beleive should happen and what I "think" I am seeing and maybe someone can tell me what is incorrect about the whole idea. The desired behavior is to have my server send an e-mail addressed to my isp or webhost e-mail addresses. In theory, this should be possible, as we do it all the time. from one server to another. Such as from my isp e-mail to my web host e-mail, which involves one server accepting a message from the other and depositing it in the mailbox of the correct user. I think I have achieved something different though. It seems that I have successfully created the message, but it is trying to deliver it THROUGH (relay?) my isp mail server instead of TO my isp mail server. Does any of this make sense? I really don't think this is such a big deal, in fact it seems like it could/should be very simple, I just have not configured something correctly. Thanks, RED

That sounds like what I want to do. I want MY server to deliver the message to my ISPs mail server so I can retrieve it from there...but how? Thanks

You just gotta love log files. It appears that my server IS working properly, but my ISP and web host mail servers are rejecting the e-mails. showing these entries: to=<my_email_address@my_isp>, relay=isp_mail_gateway[isp_mail_gateway_IP], delay=4, status=bounced (host isp_mail_gateway[isp_mail-gateway_IP] said: 550-my_IP_address blocked by blacklist.mail.ops.xxx.xxx.xxx. 550 my_ISP subscribers are no longer permitted to directly connect to this mail server. To send email to other my_ISP subscribers, you may forward messages through my_isp_smtp_mailserver (in reply to MAIL FROM command)) AND to=<my_email_address@my_webhost>, relay=my_webhost_server[my_webhost_server_IP], delay=1, status=bounced (my_webhost_server[my_webhost_server_IP] said: 550-Verification failed for <root@mylinuxbox> 550-unrouteable mail domain "my_inranet_domain" 550 Sender verify failed (in reply to RCPT TO command)) Any Ideas how to handle this one? Thanks.

I'm back! And more frustrated/comfused than ever. My question is pretty simple, (at least I think it is) but the soultion seems to be eluding me at every turn. I have a simple server set up for a friend of mine that I would like to have send informative messages to me via e-mail. this was working earlier, however the e-mails are now being returned indicating that I cannot connect directly to the mail server (my isp mail server) and that i should relay messages through their mail server (smtp.myisp.net). Well that is not really the desired behavior. What is going to be the simplest way to have the desired messages sent to me via my ISP e-mail address. I was under the impression that sendmail and postfix could handle this, however I seem to be having rouble with the configuration. I also understand the 2 mentioned apps ARE mail servers and should be able to send messages to any e-mail address. Is there some other app I should be using? Please enlighten me! Thanks in advance. RED

I just have a question and would like an opinion about a backup procedure I plan to implement for a friend's small business. Background: I have a Mandrake 10 system set up that simply offers file sharing between Windows and MAC machines for a small printing company. Thats it, nothing else. To help make the business more efficient, they are going to begin keeping active customer files/designs etc. on the server instead of constantly transferring these files to/from zip or CD. While the server is plenty capable of this, it did not have a backup device installed when built. Now that customer data will be semi permanent on this system, we need to implement a backup strategy. My friend decided that 40GB tape backups were too costly at this time so I suggested adding a second HD (80GB) to the system and utilizing tar to backup the desired files from the main drive to this drive. While I know this is not the best scenario, I figure the chances of losing both drives at the same time are pretty slim. To further limit chances of corrupting data on the second drive, it is not automatically mounted at system startup. The script first mounts the drive, runs the backup, then unmounts the drive. Any thoughts or opinions on this plan would be appreciated. Thanks, RED

Thats it!! Thanks Much! RED Marked the thread Solved - Artificial Intelligence

Hiya folks, I spent the last 2 days, with no luck, trying to find a script that will send an e-mail from a simple shell script. What I want to happen is for my simple backup script to drop an e-mail to me when it has completed. My backup is working fine from the script, but when sending the e-mail, using: "mail -s <Subject> <recipient> " it hangs waiting for output for the actual message, then with a "Cc:" prompt waiting for more keyboard input for a CC recipient then it requires a "Ctrl-D" to actually send. How can I get this all to work in one fell swoop or have the script imitate the keyboard input. Thanks, RED

Well, After trying several different configurations and possible fixes for thsi problem, I decided to take a more aggressive approach. On the macs, i donwloaded and updated their version of Outlook Express (5.0.5 I believe) and applied SP 1 and 2 to the Outlook XP (office) installation on the Windows box. This fixed the problem on the macs, but not the Windows machine. Any clues as to why or if this is related to installation of the MDK9.2 box? Thanks. RED

Hiya everyone, I deployed a MDK box for a customer as a low cost/maintenance solution for simple file sharing between Mac OS9 and Windows boxes. That portion of the box is working well, and that is all that is configured, save the normal gamut of necessary services (Webmin). There is no mail, DNS, DHCP, or web server running on this box. Their router handles DNS DHCP duties for them. My question is, is there any possible way that this simple box could break access to the customer's wail servers? Let me explain. This server replaced a W2k box a week ago, the MAC and Windows machines are swapping files back and forth all day long. The customer has Outlook Express on the Macs and Outlook XP on the windows box, checking 3 different mail addresses provided by their web site host. Everything is working AOK till about Wednesday, when I receive a call from them stating they cannot send e-mail always receiving a message "Sorry, that domain is not allowed to be realyed by this MTA" "Error 5553". A little troubleshooting, and we find that they can send mail to other domains hosted by the same company (I have one hosted by the same company, so we used it to test). But sending to other domains such as hotmail.com or comcast.net, fail with the above message. At first I thought there was some configuration that was lost but we found this EXACT same problem is happening on all of their machines. So i am thinking their router has somehow failed, but the status bar of their mail clients indicate they have connected to the outgoing mail server, it is just drop kicking the message back. To make things more interesting, i took my Win XP laptop in there and was able to send and receive mail just fine, no matter who addressed to. Now the web host insists that we do not have the "authentication required" option selected for each mail account outgoing mail server, when in fact we do. Also, all mail is sent just fine using the hosts web mail interface, so while the problem appears to be server side, our troubleshooting has proven otherwise (XP laptop and webmail successfully sending all e-mails). Is there any one thing the MDK box could be doing to cause this? I have my doubts, but it sure makes me suspicious. I have checked the configuration of the router, and disconnected the MDK box just for testing with no change in behavior. The MDK box is the ONLY network change in nearly a year. I even created a brand new e-mail account to test, still seeing the failure. Does anyone have ANY clue as to how to fix this one? Thanks in advance. RED

Thanks for the response, an I am going to guess this is the problem. To forego changing any of the default setup, I simply moved the "share" folder from /usr to / and it seems to be AOK after reconfiguring netatalk and samba. Thankd for the help. RED

Hiya folks, Well i am again having a minro issue with a 9.2 box that is serving files for macs and windows machines sharing the same physical directory on the 9.2 box. I created a folder..."share" in the usr directory and have assigned Read Write List rights to the owner (root) and the group, "group1" via Webmin File manager. "group1" has 2 users "user1" and "user2" It seems that the rights on teh /usr/share folder are changing automatically to give Read List rights to the group and changing the group to "root" as well. this obviously causes problems when trying to connect to the box froma windows or mac machine. Simply restoring the Write right to the group and changing the group to "group1' on the /usr/share folder, restores connectivity from the boxes. Any ideas what could be causing this and how do I keep it from happening Thanks, RED

I have done this from a command line, with and without the hdlist.cz location and I still get the same behavior, except i can see the command line trying all sorts of loactions for the hdlist.cz file. Thanks anyway. RED

I will give it a try, internet connection is fine. RED

Hi folks, Just installed 9.2 and attempted to update installed packages using the update tool and I am meeting with much resistance. Each time, rpmdrake fails when downloading the hdlist.cz from any/all mirrors I have tried. I then receive a message: Any ideas? RED

The current system does not have a nic and I am unsure exactly what flavor of *nix is installed. I seem to recall seeing SCO somewhere when it boots but I could be wrong. Also, it is their desire to keep this system completely off the ethernet network, they just want access from a Windows box via a terminal. RED