Two DNS servers for one domain

I had one server which acted as a dns and http server and one domain. Now I have two servers and some sites will be on a new server and the rest will still be on the old one. Is it possible to configure named on old box to ask the new one for all subdomains it doen't know about? I don't want to configure every new domain/subdomain (for sites I create on new one) on the old server's dns.

From serverfault X4lldux

• I would suggest that you actually configure both DNS servers to host ALL records. That way you have redundancy and can lose a DNS server without taking out one or more of your domains.

  The actual best practices for DNS are (not complete, but good general guidelines & starting point):

  • Have at least 3 DNS servers
  • Have them be on different network segments
  • Have them be in different physical data centers (although if 2/3 of them are in one DC and the other one is in a separate DC just make sure the lonely DNS server can handle requests for all records)
  From Zypher

• You can specify this in the named.conf I do agree with what Zypher has mentioned and redundancy is a priceless investment.

  From Nick O'Neil

Is Sql Server 2008 R2 unsupported by Operations Manager (SCOM) 2007 R2?

Hey all,

I'm performing a test configuration of System Center Operations Manager 2007 R2, on a system prepared with Sql Server 2008 R2. Unfortunately, the Scom 2007 R2 prerequisites verification program seems to be detecting exact versions of Sql Server, and not simply a minimum version, like it claims:

"System Center Operations Manager 2007 R2 requires SQL Server 2005 Standard or Enterprise Edition with SP1 and above or SQL Server 2008 Standard or Enterprise edition with SP1 and above. Note: Operations Manager 2007 R2 does not support a 32-bit Operations Manager Operations database, Reporting Server data warehouse or Audit Collection database on a 64-bit operating system."

I had hoped that this was just a helper tool that was assisting in getting me off the ground, but unfortunately it seems as if it's actually used as a gate for the installation to proceed. Has anyone encountered this? If so, is there a way to fool the installer into thinking that it has a proper version, or otherwise alert it to my valid configuration?

From serverfault bwerks

• Probably - remember SCOM 2007 R2 was released before 2008 R2.

  Best bet is to contact Microsoft PSS and ask for a hotfix. Or at leatst contact them in the official forums. Or wait - probably a service pack will come soonish.

  bwerks : Yeah, that makes sense considering Sql Server 2008 R2 was only released very recently. I'll post back when I get a response!
  From TomTom

• On the SCOM installation disk there is a application that will allow you to create the databases, regardless of the pre-reqs. You can find it at:

  :\SupportTools\\DBCreateWizard.exe

  Upon firing this up, you can choose whether you are creating a Operations Manager Database or a database for the warehouse. For me, I left everything else as the default. It created the database and I was able to return to my SCOM server and run the setup.

  From Ricardo Garrido

• Did you get it fixed? I have the same problem!

  Chris S : SCOM and SCCM do not currently support SQL 2008 R2, you must use 2005 or 2008.
  From Bobby

• Use of SQL Server 2008 R2 is currently only supported for new installations of Operations Manager 2007 R2. See http://support.microsoft.com/kb/2425714 for details.

  From Tom

What kind of SCSI terminator do I need?

I'm connecting a PowerVault 114T tape drive to a server with an Adaptec 29160 SCSI card. I know i need a terminator on the tape drive, but not sure exactly what kind I need. Do I need a multimode or SE terminator?

EDIT - Like @Chopper 3 said, the PowerVault 114T is an enclosure; the tape drive in it is a LTO-2-L(114T), it's a SCSI drive.

From serverfault jnhdny

• Roads? Where we're going, we don't need roads!

  Just google for an "active LVD SCSI terminator VHDCI male" where you live, as that's what you need for LVD and that particular drive.

  EDIT - actually given the 114T is an enclosure that can take a range of different drives within it and it's the actual drives that define what connections you have (for instance you can put SAS drives in those if you want) then it depends on what drives you have inside if you need a terminator or not and if so what type. So let us know what type of drives you've got and we'll be able to be a little more certain. Thanks.

  jnhdny : I'm hesitant to disagree with someone with such a high rating, but I'm pretty sure the interface on the tape drive is HD-68, not VHDCI; and that's where the terminator is needed, no?

  Chopper3 : Oh you shouldn't take that into concern, I'm as fallible as anyone! actually you made me go look up some more detail and as such I'm just about to make an edit :)
  From Chopper3

• Some drives have a dip switch you can set to enable termination in which case you do not need a separate terminator. (Note this is different than "enable terminator power" which is used when your HBA doesn't provide that power (most HBAs do).)

  From embobo

Two network cards, different VLANS and DHCP requests

Our network is is setup with two different VLANS, VLAN1 and VLAN2, VLAN1 is our computer system and VLAN2 our telephone system. There is no routing between the VLANS, VLAN1 has 2003 as a DHCP server and VLAN2 uses the phone system to give out IP Addresses.

We have some computers that need access to both VLANS, and we've achieved this by using two NICS and statically assigning IP addresses. We've now come to apoint where there are enough computers that we don't want to manually manage IP addreses on VLAN2.

The issue is that when set to DHCP NIC2 on VLAN2 is being assigned an IP address from the DHCP server on VLAN1. If we unplug the network card on VLAN1, then the correct IP address is being assigned, there is no routing as it's not possible to access the IP address on NIC2/VLAN2 from anywhere on VLAN1 once it has been given.

I'm curious as to why this is happening, I've been unable to find anything specific but my general feeling is that XP broadcasts DHCP requests over both network cards, even if it is only one card that is requestsing it, the 2003 DHCP server is responding on Card 1 with a higher priority, so card two is using that responce.

From serverfault Tubs

• this sounds like a bad idea to set it up in this way. I would have given the 2003 server 2 NICS and enabled the routing and remote access on it. Clients in vlan1 could then route to vlan2 via the server where we could implement some basic filtering and firewall rules.

  You Could then configured the server as a dhcp server for both vlans by adding the appropriate scopes and setting the DHCP relay agent in the routing and remote access config. this way only the server requires 2 NICS

  Tubs : Its already setup, it's been running for five years. Might be a bad idea but it works reliably.
  From MikeT

• I suspect that the PC's have the two networks 'bridged' so they are broadcasting the DHCP requests. I'd check the 'Network Connections' screen on the PC's.

  However, I agree with MikeT, there are other ways to skin this. You could use a server with two NICs, or if you really have VLANs (not just two subnets), then your networking equipment likely has a facility for setting routes between VLANs, and setting security on who/what can acccess those routes.

  Tubs : I suspected bridgeing. There isn't any.
  From BillN

403.4 won't redirect in IE7

I have a secured folder that requires SSL. I have set it up in IIS(6) to require SSL. We don't want the visitors to be greeted with the "must be secure connection" error, so I have modified the 403.4 error page to contain the following:

<script type="text/javascript">
function redirectToHttps()    {

    var httpURL = window.location.hostname+window.location.pathname;

    var httpsURL = "https://" + httpURL ;

    window.location = httpsURL ;

}
redirectToHttps();
</script>

And this solution works great for every browser, but IE7. On any other browser, if you type in http://www.mysite.com/securedfolder it will automatically redirect you to https://www.mysite.com/securedfolder with no message or anything (the intended action). But in Internet Explorer 7 ONLY it will bring up a page that says

The website declined to show this webpage
Most Likely Causes:
This website requires you to log in

This is something we don't want of course. I have verified that javascript is enabled, and the security settings have no effect, even when I set them to the lowest level I get the same error.

I'm wondering, has anyone else seen this before?

From serverfault Jeremy Morgan

• I never experience it before, but, i wonder why you use javascript to do it ? Ever tried with the ?

  The way I deal with this problem was easy, since I use ASP.NET. You make an abstract page class or HttpModule (entire web app) that check if (!Request.IsSecure) ... redirect with a 301 to secure page...

  WON'T NEVER BREAK ! Will work without javascript... I'm pretty sure you can do the same in any language (Java, PHP, ruby even c++ if you're old school).

  Wanna learn more on asp.net ? Go to Developerit.com

  Developer IT : Damn ! I got caught with my meta tag... Ever tried <meta http-equiv="REFRESH" content="0;url=http://www.the-domain-you-want-to-redirect-to.com"/>

  Developer IT : Forget it I can't paste html so ther's a link http://www.instant-web-site-tools.com/html-redirect.html
  From Developer IT

• My guess is that IE's "Friendly Error Message" feature is replacing your error page with one it thinks is easier for people to understand. According to this kb article, the "friendly error message" is not used when the server response is large enough, and suggests padding your custom error page with spaces until it is.

  From DerfK

Windows Vista/7 Registry Issue

I've been given the task of modifying a batch file for the installation of our software that sets certain registry values. The modification has to do with differences in the registry between Windows XP, and Windows Vista/7. Apparently in the newer OS's, there is a "shadow registry" of some sort that contains registry data that may be out of sync with the main one? I can't find any information on this.

Could this possibly be referring to the transactional nature of the registry in Vista/7? Would a simple "commit" resolve this? If so, what is the syntax for that?

EDIT:

By "syntax", I am referring to the syntax of doing so within a batch file? i.e. I've added keys/values, what command will "commit" the transaction?

From serverfault Ryan

• There is a virtual registry on Windows Vista / 7, it's basically for compatibility, as these OSes have changed the way that security is applied to the registry, and is per user.

  "Prior to Windows Vista, many applications were typically run by administrators. As a result, applications could freely read and write system files and registry keys. If these applications were run by a standard user, they would fail due to insufficient access. Windows Vista improves application compatibility for these users by redirecting writes to a per-user location within the user's profile. For example, if an application attempts to write to HKEY_LOCAL_MACHINE\Software\Contoso\ it will automatically redirect to HKEY_USERS\< User SID >_Classes\VirtualStore\Machine\Software\Contoso."

  And a major warning ...

  "As virtualization is an interim application compatibility technology, Microsoft intends to remove this form of virtualization from future versions of the Windows operating system as more applications are migrated to Windows Vista. As a result it is imperative your application does not take a hard dependency on the presence of virtualization in the system."

  http://msdn.microsoft.com/en-us/library/bb530198.aspx

  From Tubs

Amavis + Postfix configuration problem.

Hello we're inherited one server based on Gentoo + Postfix + Amavis-new + Maildrop

$ cat /etc/gentoo-release 
Gentoo Base System release 1.12.9
$ equery list
mail-mta/postfix-2.5.5
mail-filter/maildrop-2.0.4
mail-filter/amavisd-new-2.5.2

Some config files for postfix+amavis: http://dl.dropbox.com/u/9430338/config.tar.gz

Log snipets:

/var/log/mail.log:

Oct 15 13:36:29 myhost postgrey: action=pass, reason=triplet found, client_name=mail-yw0-f48.google.com, client_address=209.85.213.48, sender=my@gmail.com, recipient=test@myhost.it
Oct 15 13:36:33 myhost postfix/smtp[3535]: 9D576478A5B: to=<test@myhost.it>, relay=127.0.0.1[127.0.0.1]:10024, delay=3.8, delays=0.68/0/0.01/3.1, dsn=2.7.0, status=sent (250 2.7.0 Ok, discarded, id=04288-01 - SPAM)

/var/log/amavis.log:

Oct 15 13:44:47 myhost.it /usr/sbin/amavisd[4288]: (04288-05) ESMTP::10024 /var/amavis/tmp/amavis-20101015T133629-04288: <my@gmail.com> -> <test@myhost.it> SIZE=2595 Received: from myhost.it ([127.0.0.1]) by localhost (myhost.it [127.0.0.1]) (amavisd-new, port 10024) with ESMTP for <test@myhost.it>; Fri, 15 Oct 2010 13:44:47 +0200 (CEST)
Oct 15 13:44:47 myhost.it /usr/sbin/amavisd[4288]: (04288-05) Checking: I9aGeuG0kC1N [209.85.213.176] <my@gmail.com> -> <test@myhost.it>
Oct 15 13:44:49 myhost.it /usr/sbin/amavisd[4288]: (04288-05) SPAM, <my@gmail.com> -> <test@myhost.it>, Yes, score=1006.929 tag=x tag2=5 kill=10 tests=[AWL=0.390, BAYES_50=0.001, DNS_FROM_OPENWHOIS=1.13, FH_DATE_PAST_20XX=3.188, GTUBE=1000, HTML_MESSAGE=0.001, TVD_SPACE_RATIO=2.219], autolearn=no
Oct 15 13:44:49 myhost.it /usr/sbin/amavisd[4288]: (04288-05) Blocked SPAM, [209.85.213.176] [209.85.213.176] <my@gmail.com> -> <test@myhost.it>, Message-ID: <AANLkTi=UyxaEAOG6+qXQfx+Pbm4if+Hs9BBqyoQHtBR7@mail.gmail.com>, mail_id: I9aGeuG0kC1N, Hits: 1006.929, size: 2595, 2006 ms

PROBLEM.

The problem is that in current configuration the mail system drop all spam message.

Desired behaviour.

We would like to receive all messages in the IMAP folder .Junk with ***SPAM*** message prepended to the subject.

Thank for your time!

From serverfault Igor

• What value do you have set for "$final_spam_destiny" (amavisd-new)? It should be D_PASS;

  Sorry if it's in your config file, I didn't look at it yet.

  Igor : YES you may check in amavisd.conf: $final_spam_destiny = D_PASS;, but still no result. It seems that amavis actually passes the message back to postfix and postfix drops it.
  From Mark