bakytn's questions

My simple goal is to have two instances. one is for database, second for just a site. Undoubtedly those 2 instances must have minimum network latency to communicate.

I know that there are so called placement groups but the problem is, they require expensive instance types to use.

Another option (to my mind) would be to use VPC. However, I can't use them. For some reasons in Management Console I can't create any VPC (look at the screenshot).

So my question is.

If I just create instances in the same region, would it work? I.e would they have a LAN like speed?

I have a samba server and just installed the second one to configure a reserve file server.

It all went well and both of them successfully are on Active Directory domain (actually SAMBA 4)

the problem is that on the server 2 user shares messed with wrong uid/gids.

and if I id user1 on server 1 I get 2010 but but I run id user1 on server 2 I get 2031

They use 99% identical smb.conf

and they have: passdb backend = tdbsam

I have found out that in order to achieve equal UID/GID's I would need to use ldapsam but I don't know how to do it.

Do I need to first export from tdbsam and then install ldap server on server 1 export to that ldap server?

and then set passdb backend to ldapsam on both?

Can you help me with understanding of reliability of network adapters.

Most of the time servers do have at least 2 NIC's bonded to provide sort of a HA for it. So in case of one NIC fails, the second would still do the job.

I wonder which factors work when you use network adapters.

I know that, the most important and weakest part of any computer system is: storage (i.e HDD).

but how reliable actually network adapters are?

There are more expensive ones, and cheaper adapters.

In which cases do they actually fail? In what circumstances.

1. May it be a intensive usage of them
2. Time when it's on

In your experience how often you found yourself changing NIC's due to their fail?

Or just what's the typical lifetime of commodity NIC's?

thanks.

According to http://en.wikipedia.org/wiki/IOPS

For example SAS disks provide ~200 IOPS each.

For RAID 10 consisting of 4 such disks.

How much IOPS I could expect for that array?

I would think it would be about 400 or is it 800? (800 especially in reads? and lower on writes?)

Today I found out that I don't know how to fix IP conflicts.

Say you have a server with IP1

another machine connects and assigns itself IP1.

How do people usually prevent this? How they(you) resolve?

I heard that usually it is secured by ROUTER. Is that correct?

And it is the only way?

UPDATE:

My question was about the situations when some user just assigns server IP to itself and that would cause problems.

But I found out that system administrators just should control all the hosts and machines and never allow users to assign any IP but given by Admin (read as given by DHCP server)

I have a Server with 5 NIC's.

And I have just installed Ubuntu Server 11.10 64 bit fully dist-upgraded by now (31 Jan)

3 NIC's are identical PCI-Express x4 adapters. 1 is integrated TG3 and the last one is PCI but with the same RealTek chipset as those 3 NIC's (r8169)

My system is only seeing three NICs and assigning their eth* other two adapters are just not visible. It can't see TWO of those THREE identical adapters.

I also couldn't find them on DMESG.

Adapters by themselves work fine I also tried to plug the cable and their LED's blinking (I also checked them on another machine and they work fine).

Please help me!

What can I do? How can I debug? I need at least one of them to function properly.

My server supports hardware RAID and it can use SATA drives.

I wonder does it depend on Server or on a Discs to know, if they support hot swapping? And rebuilding while working?

I know that SAS drives are intended to be used as a hotswap. How about SATA's?

My friend accidently bought Japanese MS SQL Server 2008 R2 instead of the English version.

However, we currently have English Windows Server 2008 R2. Can we install it anyways?

I don't have a test machine to try it out on first to see what happens.

Can I deny "Internet Explorer" via GPO and use Google Chrome as a default and the only Browser?

I wonder as Internet Explorer is a part of an Windows operating system, if it's possible to deny it.

I would like to force users to use Google Chrome only as a browser.

I have iptables and masquerading enabled through this:

iptables -t nat -A POSTROUTING -o eth0 -j MASQUERADE

How can I revert this rule? i.e I want to disable the masquerade.

I am afraid to test..because if I fail..there would be a lot of problems (for example lose ssh)

Also can I restart iptables hoping that it wouldn't start rejecting all conns.

I was "debugging" my Folder Redirection GPO

What would happen if I uncheck "Exclusive rights to owner" in Folder Redirection GPO Editor?

I'm interested in consequences. What would happen to the current user permissions?

Would it be applied to their folders as well? Or it works for new users only?

If you are interested, I am debugging this: https://serverfault.com/questions/327312/folder-redirection-gpo-is-not-working-for-certain-users-gpupdate-doesnt-work

I have set up ejabberd now the problem is.. (look at the screenshot). Users complaining about it it confuse them.

How can we hide the number?

I was playing with Folder Redirection feature of Microsoft.

the only domain controller is under: Windows Server 2008 R2 the workstation is: Windows 7 Enterprise.

I have set GPO, Folder Redirection for Documents only.

Also, I didn't set profile roaming.

Now, when I log in to the Workstation, I see (in Properties) that Redirections are really working. However it's using TEMP profile anyways, even though I set FR only for Documents. not for Desktop nor any other folder I also didn't set up Profile Roaming.

WHy is it showing me this? I am missing something?

I have set up Terminal Services in Windows 2003. Which is part of the domain. That server is not DC. It has only one role: Terminal Server

When I'm trying to log in to it via RDP as an Domain Administrator or any other user in the Domain Admins group. It's working fine and allowing me to work.

However when I'm trying to RDP in with members of "Domain Users" group

I am getting:

Remote Desktop Connection Denied because the user account is not authorized for remote login

I have set local policies and allowed Domain Users (Allow logon through Remote Desktop Services) and the Deny login through Remote Desktop Services is not configured.

I also tried to add certain users to the Builtin Remote Desktop Users group no luck.

Please help!

My server uses SAS 15k RPM disks. One of them has a red light. AFAIK, that means that it needs to be replaced; but it's not used intensively - actually the server was still in the configuration stage, and only 2 months old.

What can it be? May be it's S.M.A.R.T failure? Also, when I was away the workers hard reset the server two times.

(resolved: AppArmor was the reason of that failure. I just run: aa-logprof and restarted bind)

I added second interface to my server

and after I restarted my server I have found out that BIND is not starting.

If I try: sudo /etc/init.d/bind9 restart I'm. getting

 * Stopping domain name service... bind9
    rndc: connect failed: 127.0.0.1#953: connection refused
                                                               [ OK ]
 * Starting domain name service... bind9
                                                               [fail]

contents from /var/log/syslog:

Aug 12 15:50:36 server1 named[1127]: configuring TKEY: failure
Aug 12 15:50:36 server1 named[1127]: loading configuration: failure
Aug 12 15:50:36 server1 named[1127]: exiting (due to fatal error)
Aug 12 15:50:36 server1 kernel: [   38.746404] type=1400 audit(1313142636.718:15): apparmor="DENIED" operation="file_lock" parent=1126 profile="/usr/sbin/named" name="/usr/local/samba/private/dns.keytab" pid=1128 comm="named" requested_mask="k" denied_mask="k" fsuid=103 ouid=103

However I think the problem is not coming from the second interface as I removed it and rebooted. Having the same issue.

I am setting Windows Server 2008 R2 as an Application Server (RDP, SeamlessRDP). I was going to first install that app on the server and then provisioning it to the clients.

The problem is that it's 32 bit app and when I'm trying to install it I'm getting:

The version of this file is not compatible with the version of Windows you're running. Check your computer's system information to see whether you need an x86 (32-bit) or x64 (64-bit) version of the program, and then contact the software publisher.

I tried to set compatibility mode with XP SP2 and SP3, tried Windows 2000. Btw, It works perfect on XP SP2, SP3.

The installer is InstallShield. The Windows built in compatibility wizard is didn't work. I played with other settings like "Run as Admin.." and work in 640x480 mode - no results.

Other apps (also 32bit) are working fine

After my ltsp-build-client --arch i386 All went well, but for some reasons tftpd-hpa wasn't installed there were no /var/lib/tftpboot folder.

So I had to install tftpd-hpa manually. DHCP is working well but it can't find pxelinux.0 file (it just doesn't exist).

Here is my DHCP server conf.

authoritative;

subnet 10.10.10.0 netmask 255.255.255.0 {
    range 10.10.10.20 10.10.10.250;
    option domain-name "example.net";
    option domain-name-servers 10.10.10.1;
    option broadcast-address 10.10.10.255;
    option routers 10.10.10.1;
    option subnet-mask 255.255.255.0;
    option root-path "/opt/ltsp/i386";
    if substring( option vendor-class-identifier, 0, 9 ) = "PXEClient" {
        filename "/ltsp/i386/pxelinux.0";
    } else {
        filename "/ltsp/i386/nbi.img";
    }
}

client machines are saying:

PXE-T01: File not found
PXE-E3B: TFTP Error - File Not Found

PXE-MOF: Exiting Intel PXE ROM.

The question: How to generate the pxelinux.0 file and/or nbi.img file? There is image though in /opt/ltsp/i386/images/i386.img But I'm not sure what's that image for.

TFTP Server is running by: /usr/sbin/in.tftpd --listen --user tftp --address 0.0.0.0:69 --secure /var/lib/tftpboot

the /var/lib/tftpboot folder is empty.

Or another question. How to Generate pxelinux.0 and nbi.img file again?

We use local Ubuntu repository mirrors, because our external traffic is not free. Whenever I apt-get install "program" it gets from that repository.

The question is... can the maintainers of the repository substitute any package on the repository with their own package?

Can I be hacked easily on any apt-get upgrade or apt-get install or apt-get dist-upgrade?

We get very basic packages from the local Ubuntu mirrors, such as "telnet" or any other.

I have a PC where XenServer 5.6.1 fp-1 has been successfully installed.

I've manually set the network settings:

192.168.1.50
255.255.255.0
192.168.1.1

but it's set to xenbr0 iface. While eth0 is empty.

When I click on "Configure Management Inteface" it shows that eth0 is connected.

But when I ping a default gateway (which is 100% should be accessible) it fails.

I used to another shell (Alt+F3) and logged as root. I also failed to ping.

with both: ping -I eth0 192.168.1.1 and ping -I xenbr0 192.168.1.1

Be assured that:

1. Cable works
2. Ethernet adapter is 100% functional (prev OS was Ubuntu it was working)
3. There is no firewall rule to deny anything. (everything is allowed)

So the question is:

What is a problem???