• i7 2600K
• Gigabyte P67A-UD4 B3
• 8GB DDR3-1866hz 9-9-9
• 1TB Samsung F3 drives
• Intel Gigabit Network card

Now, the issue here is that after a clean install of VMWare Sphere 5, boot-up would take at least 5-9 min to complete from a clean power-on. Compared to an upgrade from VMWare Sphere 4.1U1 which will load all the necessary drivers right away after a restart/power-on. It seems like its not detecting the HDD at all at first glance but soon found its way. Further tests, experiments and research yielded that this is a side-effect on how VMWare handled their installation process in VMWare Sphere 5 compared to 4.1U1. In 4.1U1, it uses the MBR for the boot process by default, while in 5.0 it uses GPT. You can fixed this if you force the installer to use MBR during the installation phase.

A developer from VMWare confirmed this and posted a way to fix this. You can find the thread here.

To quote him…

By default ESXi 5.0 uses GPT based partitions tables and not MBR like it did previously. This allows you to have disks which are greater than 2TB in size, but some systems are incompatible with it (possibly including yours) or need to have the correct BIOS setting.

You can try changing your BIOS to see if it’s compatible with EFI or will work in “legacy/compatibility mode”. Alternatively you can reinstall and use the “formatwithmbr” setting at the boot prompt which will cause the partitions to be set up the way they were in ESX 4.x. Just add it in the same way you would to tell the machine to do a scripted install.

To make things short, as written by wdbarker….

the boot prompt appears after you hit Shift-O during the boot process. You will see runweasel already in the boot prompt. Right arrow to the end, add a space and formatwithmbr and press enter. The prompt will look like: runweasel formatwithmbr

So, there you go. Hope this helps.

This was in collaboration with a friend of mine (Earlz) for setting up the box and helping me debug.

sudo aptitude install acpi acpi-support apm

This will install the power-button-power-off feature in you server.

1. Download the driver source from Intel driver webpage here, or the latest one they have. Double check if the driver supports the version of the network card, the one I have is 82579.
2. Ready the Ubuntu server, install make gcc g++ and linux headers

   aptitude install make gcc g++ linux-headers-`uname -r`

3. Un-pack the src driver that you just downloaded, compile and install.

   gunzip [filename].tar.gz
   tar -xvf [filename].tar
   cd [folder_name]/src
   make install

4. If everything goes well, no error should come out, then enable your new ethernet card.

   modprobe e1000e

5. You can check your new network card via ifconfig
   Enjoy!

Note: If you update your kernel or linux image via aptitude safe-upgrade or dist-upgrade, you need to remove the e1000e module, re-compile the driver and install the newly compiled driver again.

In your Shorewall Rules file, add this above the other rules.


DNS/ACCEPT  $FW     net
DNS/ACCEPT  net     $FW
DNS/DROP    loc     net
DNS/DROP    net     loc

$FW is your firewall/gateway and in my setup my DNS. You can change $FW to loc:<dns_ip_address> if your DNS is located in another machine other than your gateway.

Have fun raining in on their parade

P.S. You can add filters to allow certain IP or MAC addresses to use DNS outside (i.e. Google DNS ).

• Any latest Ubuntu server ( I’m using Ubuntu 10.10 Server )
• ddclient installed (use “aptitude install ddclient”)
• net connection, I have set up my linux server as a firewall/gateway

After installing ddclient, go to /etc and edit ddclient.conf, below is the format that I used, modify it to suite your configuration.

ssl=yes
use=web, web=whatismyip.org
server=updates.opendns.com
protocol=dyndns2
login=<openDNS_username_here>
password='<openDNS_password_here>'
<openDNS_network_label_here> #this can be found under settings, just the label, nothing else should be added here

Save the new configuration file and restart the client /etc/init.d/ddclient restart
Also make sure that the network that your going to update the IP address has set its Dynamic IP Update option enabled. This can be found under the advance setting of the network.

The catch?
XML does not support named entities by default, you have to declare it before hand. Only &lt; &gt; &amp; &quot; and &apos; are predefined.

Something like:

1
2
3

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE xsl:stylesheet [ <!ENTITY nbsp "&#160;"> ]>
<xsl:stylesheet xmlns="http://www.w3.org/1999/XSL/Transform" version="1.0">

Got this from here…

The fix…
Instead of using named entities in the XML (I think this also affects JSON, I could be wrong since I didn’t test this is JSON) like &ntilde;, use character references to support special character encoding.

Example:
white space = &nbsp; = &#160;
ñ = &ntilde; = &#241;

they are all the same… #160 is referenced/mapped to ” ” and #241 is referenced/mapped to ñ

So when forming your XML before returning it through an AJAX call, form the special characters via character reference.

Hope this helps you out there having this problem.

So anyone having this problem, check the following…

• What OS are you running? Vista/Win7?
• What IDE are you using?
• Was your IDE elevated to admin status?

So, check your OS, I’m using Win7. I was developing this under WinXP a few months back, and now under Win7, the folder is protected since it was under the “Program Files (x86)” folder group. (Argh!)
I closed my IDE, using IDEA IntelliJ 9.0, ran the IDE under “Administrator Level” and the problem disappeared!

So check those out first.

I was following the tutorial from Sun in creating my first JNI and calling my compiled C++ library, I was stuck in the last part where this error popped out whenever I tried to invoke

java HelloWorld

the error was…

java.lang.UnsatisfiedLinkError: no HelloWorld in java.library.path
          at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1709)
          at java.lang.Runtime.loadLibrary0(Runtime.java:823)
          at java.lang.System.loadLibrary(System.java:1030)
          at HelloWorld.<clinit>(HelloWorld.java:<line number>)
Could not find the main class: HelloWorld. Program will exit.

one way of fixing this is to include the current location of your newly compiled shared library (ex. HelloWorld.so). I’m using Ubuntu linux,so to include the current folder where your “.so” file is located, invoke…

LD_LIBRARY_PATH=`pwd`
export LD_LIBRARY_PATH

or you can also do …

LD_LIBRARY_PATH=.
export LD_LIBRARY_PATH

another way is to define the location when running the program…

java -Djava.library.path=. HelloWorld

this will instruct java to locate the native library in the current location. Now call your program again and this time it work execute your native program.

It’s possible that when you execute your program after doing the above steps, another error will come out. This happened to me and the culprit was when I compiled the cpp file. The error would look something like…

java.lang.UnsatisfiedLinkError: <library location>: <library location>: only ET_DYN and ET_EXEC can be loaded
        at java.lang.ClassLoader$NativeLibrary.load(Native Method)
        at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1778)
        at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1674)
        at java.lang.Runtime.load0(Runtime.java:770)
        at java.lang.System.load(System.java:1005)
        at HelloWorld.<clinit>(HelloWorld.java:16)

to fix this, include the “-shared” option when compiling the cpp file.

g++ -shared HelloWorld.cpp -o HelloWorld.so -I /path/to/your/java/include/folder/where/jni.h/is/located -I /path/to/your/java/include/linux/folder/

Do not include the “-c” option, this will skip the linking stage.

That didn’t fixed the problem.

There was something else running when windows starts up. Next target… DLLs. Amazingly, there were dlls that was out-of-placed and curiously named. Example crypt32.dll that wasn’t signed by microsoft. Removed the entry and tried to remove the dll. No luck, its being used. Next tool that was very useful was Unlocker. I ‘unlocked’ the hook for the dll and restarted. Still didn’t fix it but I was getting close. There were two more dlls I found that was very suspicious. I tried to kill one and removed, the other regenerated the dll. I try to unhook the other dll, the whole pc restarts. Smart bugger. Thank goodness for safe mode. Restarted in safe mode and use unlocker again to unhook the dll. Removed one of the dll and its entry in the registry. Good… it didn’t restart… just a countdown for a system shutdown.. freaking dll.

The system restarted and good news is the dll that forced the shutdown wasn’t regenerated. Last dll was a bit simpler. Used unlocker to unhook the dll, deleting was not successful, though unlocker had a feature to delete the dll after a restart. Selected the option and waited. Thankfully, the whole ordeal ended there. All three dlls didn’t regenerate after that.

Note to PC users who watch videos from torrent ( Heroes, Prison break…). If it asks you to download a player to run the video you downloaded.. DON’T!!

FLUSH LOGS;

Bring in a fresh copy of the database from the master-server and update the slave-server database. THIS IS IMPORTANT! Since if you don’t update the slave database, you will not have the data from the time you were disconnected until you reset the relay logs. So UPDATE YOUR SLAVE WITH THE LATEST DATABASE FROM THE MASTER!

Now when the logs are flushed,all the relay-bin logs will be deleted when the slave is started again. Usually, this fixes the problem, but when you start the slave and the failed relay log error is still there, now you have to do some more desperate measures… reset the slave. This is what I had to do to fully restore my MySQL slave server. Reseting the slaves restores all the settings to default… password, username, relay-log, port, table to replicate, etc… So better to have a copy of your settings first before actually do a slave reset. When your ready to rest the slave, do the command…

RESET SLAVE;

after which you should restore all your setting with a command something like…

CHANGE MASTER TO MASTER_HOST=.....

now start your server with…

SLAVE START;

check your slave server with…

SHOW SLAVE STATUS\G

look for …

Slave_IO_Running: Yes
Slave_SQL_Running: Yes

both should be YES, if not, check your syslog if there are other errors encountered. I’ll leave this until here since this is what I encountered and I was able to fix it.

Edit 5/14/11:

There is a possible change that after executing the CHANGE MASTER command that you’ll receive this error below…

ERROR 1201 (HY000): Could not initialize master info structure; more error messages can be found in the MySQL error log

This can occur when the relay logs under /var/lib/mysql were not properly cleaned and are still there. The next thing is to delete them manually, log back in to mysql, refresh logs, reset slave then execute the CHANGE MASTER command again. The file to delete would be relay-log.info .This should work now. Sometimes I don’t know why mysql can’t reset the slave logs.