Mike B's questions

I have an Apache/2.4.6 installation working as a reverse proxy for some applications. Now I have to configure another one. HTTP proxying is working great, but in some sections of the webpage, the client tries to start a websocket connection.

The problem is that the url is partially shared:

HTTP request: http://domain/context/[whatever]

WS request: ws://domain/context/specific-text?param1&param2

Cannot use If (checking the protocol) as ProxyPass cannot go inside <If>.

Tried setting:

<Location /context/specific-text>
    ProxyPass ws://internal.domain/context/specific-text

Before

<Location /context/>
    ProxyPass http://internal.domain/context/

But doesn't work, guess because of the (non existant) trailing slash.

More data:

• The socket is complaining about not liking a "200" code, Apache is not returning a ws conection but, to some point, the proxying is being made.
• Tried with and without <Location>
• Thought of rewriting specific-text to specific-text/ but I don't like to manipulate things I don't manage. The application is a paid product, can't ask if it would harm anything. Well, I could, I'd just wouldn't expect any answer.

We're serving mail for a group of customers, and (only) one of them is seeing (some of) their mails rejected.

The situation is:

• Our server: ourdomain.com -> Gets forward and reverse DNS lookup correctly resolved
• Served domain: clientdomain.com

Dig over clientdomain.com:

clientdomain.com.       2742    IN      MX      5 mail.clientdomain.com.
mail.clientdomain.com.  2742    IN      A       ip.of.ourdomain.com

Error to mail clients:

Recipient address rejected:
SPF-Result=ourdomain.com: 'SERVFAIL' error on DNS 'SPF' lookup of 'ourdomain.com' (in reply to RCPT TO command)

My main question is...

Would it be better something like

clientdomain.com.         2742    IN      MX      5 mail.clientdomain.com
mail.clientdomain.com.    2742    IN      CNAME   ourserver.com

?

Secondary question... if above answer is NO, what can I do? add an SPF record in favour of ourserver.com? Why aren't other clients noticing this?

Thanks for reading

I'm trying to configure RAID in a IBM x3550 M2 server. I can't find any messages on boot about any hotkey to access the configuration utility. I wasn't able either in BIOS.

The card is a on board LSI 1068e. I can't find the original CD's shipped with the server, but I downloaded Server Guide 9.21 from IBM web. Booted with that, and only getting stuck at "Windows loading"p. Tried with a 8.? version from a x3550 (M1), but it was less useful.

Although I do not consider myself an expert, I'm starting to feel like a newbie luser. Any help?

I've been searching for similar questions, but couldn't find any useful, sorry if duplicate. (I read windows 2003 server slow access and Horrible Windows Server 2003 performance while copying files through Netbios)

We've got a ML 115 G5 server with Windows 2003 32bit. It acts as file server, LAMP server and Trend Micro Office Scan server. It also has an old MSDOS program.

It's less than 2 years old. In the 100mbps network there are only two clients. One of them uses more often the MSDOS problem, and the LAMP server, which run more or less OK. The problem is with the other client. She opens Office documents and this is where the server comes to a worrying slow speed.

We run an HD Tune test, and we found 28ms of latency and slow transfer ratio. We updated to the latest NVidia Raid software driver (RAID1, by the way), and it improved just a little. We went down to 23ms of latency and better transfer ratio, but still low.

It's not a matter of the switch, as it is slow even with a direct crossed cable connection. Tried with small files and a big iso file.

Tried also disabling antivirus client, as most of the hard disk IO was Trend Micro related, but still the same.

Windows is fully updated, as well as the network card drivers. Memory is less than 50% occupied, and none of the 4 CPU reach 20% of proccessing.

My main question is, what can I look at or where can I search to find the bottleneck?

I set up a 2008R2 AD in a small office (less than 10 users). We are opening a delegation, and I'm studying the different choices for the domain (let's say domain.com). The link between the offices will be a WAN accelerator with an IPSec tunnel, as both of them are connected to Internet through ADSL connection (around 6Mbps/600Kbps). File server and ERP application will remain in the main office.

I've been searching for a comparison chart or a cons-pros list, but I haven't found anything.

First choice, secondary domain. AFAIK, it would imply the creation of a subdomain (branch1.domain.com). Would it also imply that users are different or anyone could log in any of the domains? I guess DNS management would be independent between zones. Any other consideration?

A domain replica is another option, but I'm not very expert, and I fear DNS management of the different offices. On pros, I guess it would speed up login and would make my domain high-availability. But another office will probably be opened, any problem with 3 replicas? I read that replication is made in a ring manner, but links are only point to point with main office, I don't need offices to be seen between them. Would this be a problem?

The third option, a Read Only Domain Controller seems to be the best, as it would have the good thing about replication with no (or less) considerations. It will also have the bad thing about DNS management.

If someone could lighten this up, or point me to a tutorial or something where this things are explained, I would be grateful. I've been reading about the three options, but can't find any comparison between them.

I have a 1TB USB disk where I make the first level of backup of a WS2008 R2 file server. The usb is not fully encrypted, just about 700GB. Cobian is running as a service, impersonating Administrator user when doing backups.

It happened that "accidentally" some guy unplugged the USB and then plugged it again. It is not a big deal as everyday I check the backups, but I wanted to make a script to mount the encrypted file just in case before doing the backup. The script works ok if I execute it, but it seems to fail if executed by Cobian with no session opened. Cobian always fails as doesn't find the mounted unit (T:). The script, in a .bat file:

"[fullpath]\truecrypt.exe" /v z:\backups /l t /k [keyfilepath] /a /q /s /c n > "C:\mount.log"

Z is the usb, it always takes that letter when plugging it. Also, mount.log is always empty. I've read the TrueCrypt Known Issues and FAQ and I can't see anything related. Any idea about why does it fail or how to get log messages?

PS: I also tried mounting the encrypted file as Removable Media, with no luck. The USB gets mounted itself as drive, not as removable media.

I'm building a replica of a production server for migrating it to a Virtual environmnet. I do not want to just copy it as it is a 32 bit CentOS 5, and I upgraded it to CentOS 6 64bit.

I managed to replicate mysql and maintain it as a slave of the production one. The mailserver is managed through ISPConfig2, which I also copied.

The problem now is authentication with pop server. I installed auth-mysql library, and configure it (comparing the configuration files in the production server). I managed to get it loaded on startup, but it seems to be another configuration overriding it:

Sep  5 09:27:10 localhost authdaemond: modules="authmysql", daemons=5
Sep  5 09:27:10 localhost authdaemond: Installing libauthmysql
Sep  5 09:27:11 localhost authdaemond: Installation complete: authmysql
Sep  5 09:27:18 localhost postfix/postfix-script[1502]: starting the Postfix mail system
Sep  5 09:27:18 localhost postfix/master[1503]: daemon started -- version 2.6.6, configuration /etc/postfix
Sep  5 09:27:27 localhost authdaemond: modules="authuserdb authldap authcustom authpipe", daemons=5

Could it be postfix overriding it? It first loads "authmysql", and after that, starts loading "authuserdb authldap authcustom authpipe". I've been trying all the weekend, I even removed auth-ldap configuration files, but I'm now stuck.

When I login via Telnet, it even doesn't try to connect MySQL:

Sep  5 09:41:00 localhost pop3d: Connection, ip=[::1]
Sep  5 09:41:09 localhost authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server
Sep  5 09:41:09 localhost authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server
Sep  5 09:41:09 localhost pop3d: LOGIN FAILED, [email protected], ip=[::1]
Sep  5 09:41:09 localhost pop3d: authentication error: Input/output error

Any idea on what to try or which configuration file could be doing this would be appreciated.

I have a pfSense Firewall with OpenVPN against RADIUS Server configured in a Windows Server 2008 R2.

The RADIUS server is configured in pfSense, but when I try the Authentication (Diagnostics -> Authentication) I always get an Authentication Failed error.

In Windows event viewer I get a message, identifying correctly the user trying to log (domain\user), but saying "Code 16, User does not exist or incorrect password". I've tried with a couple of users, one of them Administrator.

The network policy says any user in the domain can log. I have no idea to keep trying. Any hint?