Open Source Web Hosting

While researching the hardware to purchase for a new VMWare server, I found that this SuperMicro X8DTL-3 Motherboard supports ESXi4, according to Supermicro’s web page. So I assumed it would work the the later version of ESXi (5).
However, upon receiving the server and trying to install ESXi 5, there was no local storage available with 2 SAS drives hooked up to the LSI Megaraid controller (black SATA ports).
I was able to plug in a SATA drive to the blue ports and then ESXi could see that as local storage.

Update: So the Supermicro site has a note (F) that this ESXi 4 compatibility is with “IT Mode”, and to get IT Mode the PDF manual says to hit ctrl-c in the bios, so i tried that (did it do anything?) booting up the installer again now… still no new local storage (just the SATA drive is listed, where are the SAS drives?)

Update 2: the PDF manual also says to “flash an IT mode firmware to the EEPROM” – I wrote to supermicro support and they sent me this link to download “an IT mode firmware” http://www.4shared.com/zip/CDEtiP5c/X8DTL3_IT_112310.html which I will now attempt to flash.

Update 3: I had to put those files onto a bootable USB stick, now trying the instructions:

Software RAID Flash Instruction.
================================

2) Flash with BAT file: (Automatic flash)
a:\>***.BAT
(It will flash Firmware, BIOS and SAS address; but please manually type
in the address)

So I do these 2 steps (with the reboot in between) and it asks me to enter the full 64-bit SAS address. Another Email to support later and I find out that’s on a white sticker on the Mother board. Gratefully Supermicro support gets back to me with real answers in an hour or two. Frustrated that even after that, still no SAS drives showing up anywhere…

Update 4: I found in the manual there is also a physical jumper on the motherboard, after opening that the SAS drives are visible to the bios as potential boot drives. I also had to set the SATA bios settings to compatible / enhanced so that both the SAS and SATA drives are visible. Looks good, on to installing ESXi.
Looks like ESXi 5 is going to work, at least it’s seeing the SAS drives! I only lost about 1 day figuring this out. I am so happy it worked out!!

with VMware vCenter Converter Standalone, I’m trying to clone (P2V) a running virtual Windows XP computer, but when I enter the Administrator’s username and password, I get the error as follows:

the operation could not be completed for '192.168.1.2\ADMIN$' due to incorrect user credentials

[thanks to brewgoat] First of all, make sure you’re trying to log in the way converter expects (this seems to be the biggest cause of the issue)

In our network we use username@domain.lan format but converter hates that. It requires the use of domain\username instead. If you use the wrong credential format it will give this error.

In my case, turned out IIS service was running on the source XP computer. So I just had to figure out how to disable that (manage computer, services, find IIS Admin and click ‘disable’) and then the converter started right up.

While trying to do a VMware Physical to Virtual conversion, the task failed at 1% with an error mentioning “Crypto Exception”. I found the solution here:

“In Windows Explorer, browse to the folder C:\ProgramData\VMware\VMware vCenter Converter Standalone\ssl
When you try to open folder ssl, you will discover that you do not have permission to do so.
Assuming you have administrator privileges:
When you double click to open folder ssl, a pop-up will present a CONTINUE button which, when pressed, will permanently unlock the folder for you and the problem will be solved.”

This seems to have worked as my task is now well past 1% and continuing without an error 🙂

I was trying to connect from a NAT’d server out to an FTP server on the internet. The initial login would go fine, but then when I try to do anything else,

550 Data connection must go to same host as control connection.
ftp: bind: Address already in use


After some hair-ripping, the solution was easy enough. Type the word “passive” after initially logging in. Problem solved.

I cloned an Ubuntu Virtual Machine, but the networking wouldn’t start;
/etc/networking restart led to:
“failed to bring up eth0”

A lot of people have the answer to this one, this guy was nice and concise
cd /etc/udev/rules.d/
mv z25_persistent-net.rules z25_persistent-net.rules.old (in my case the name was 70-persistent-net.rules)
reboot
voila

I needed to enable root ssh login on an Ubuntu (VM) server. Yes I know it’s a security risk, blah, blah, blah, this is a temporary setup so I can clone a VM. I googled around and everything said to simply set the root password and make sure to have “PermitRootLogin yes” in sshd_config. But try as I might, after making those changes and restarting ssh, I just could not log in. I tried rebooting the server and still no dice. Googling indicated others had been in this predicament, but no solution could be found….
Finally, I re-installed openssh-server with “apt-get install openssh-server” and then bingo it worked.

Posting this here so that next time it happens and I am saying to myself, “what the heck did i do last time this happened?”, I can find this and remember!

Edit: Upon further investigation, it seems that Ubuntu’s default ssh server configuration is in here:
/usr/etc/sshd_config
rather than the one found in /etc/sshd_config that we’re used to seeing.

I’ve been running fail2ban for a while but it wasn’t catching everything hitting the mail server. I kept seeing these guys with entries like this:
vchkpw-smtp: null password given admin:75.146.111.10
who weren’t getting blocked even though they are hitting the server over and over.

So first i created a test file, maillog.11 with the entries i was interested in, like so:
cd /var/log
grep “null password given” maillog.1 > maillog.11

Then I did some modifying and testing with my fail2ban regex files like this:
fail2ban-regex /var/log/maillog.11 /etc/fail2ban/filter.d/vpopmail.conf

After trying a few regex lines, I found the regex rule needed:
vchkpw-smtp: null password given [^:]*:<HOST>

So that takes care of those clowns, now to look at cutting down on the constant spam bombardment attempts.

We run a cool little spam filter called spamdyke. It makes log entries that look like this:
Oct 13 04:01:33 SERVER spamdyke[2978]: DENIED_RBL_MATCH from: marin@omniweb.com to: marin@omniweb.com origin_ip: 119.154.186.78 origin_rdns: (unknown) auth: (unknown) encryption: (none)

I noticed that the same ip addresses try to spam our server (and get blocked by spamdyke) over and over. GO AWAY ALREADY! OK, so more regex experiments and the gate has been closed on repeat spammers!
Here’s my 3 line regex, the last line now blocks these spamaholics (note the space between the colon and the host sandwich)
failregex = vchkpw-pop3: vpopmail user not found .*@:<HOST>
            vchkpw-smtp: null password given [^:]*:<HOST>
            DENIED_RBL_MATCH from: [^@]*@[^@]*@[^:]*: <HOST>

restarted fail2ban and it blocked a bunch of new ips. ahhhh 🙂

I have had an android phone for about a year. Before that I had an iphone. But switched after hearing about how much better android phones are. Since then I have severely regretted the decision in almost every aspect except that I am no longer paying AT&T for service.

Admittedly, my phone is a cheap model (HTC Hero) but my iphone which I bought 2 years earlier was no more expensive than this hunk of junk.

Here are my top and most recent complaints.

1) Making calls can take a long time (is this a factor of my network?) Sometimes when I press “call” nothing happens for 10 seconds or more, so I hit “call” again and end up with 2 live calls. This never happened on iphone. This sucks.

2) When I look on the map and want to see ‘my location’, 9 times out of 10 i get the message ‘your current location is not available’. If I reboot the phone (see 3) then the location works after the reboot. What’s up with that?

3) rebooting the phone takes about 4 minutes. What is this, windows 3.1? my cheapy iphone takes about half the time.

4) if i type in a number but don’t hit “call”, by the time i walk outside, the phone has gone into “sleep” mode. Fine, but the number I typed in is gone. Now I have to go back inside and look up the number again so I can go outside and make a call. I don’t want to store this number permanently, just keep it in the buffer when you go to sleep mode, phone!

ugh… looking forward to getting back on the iphone full time…. sorry android people.

I’ve installed 5.0 a few times now and I seem to remember this coming up before.
While installing it goes to a screen showing 34% progress and just stays there for a while.
During which time I start wondering all kinds of things; Did it break? Should I hit ctrl + alt + del? Should I just wait a little longer? STILL 34%! Should I go to the bathroom and come back? yes? OK when I get back, it says “installation complete” woo hoo!
So it seems like that’s the trick – if your installation gets stuck at 34%, go to the bathroom and come back.

As seems typical with everything java.net so far, instructions are wrong, links outdated, and following ‘how-to’s doesn’t quite get you where you’re trying to go.

While trying to do a SVN checkout of the hdcookbook stuff, i followed about 5 bad links before finding this post.

Here is a current working repo url:
https://svn.java.net/svn/hdcookbook~svn/trunk