Tag Archives: unix

Solaris 10 in a VM

I had the chance today to assist a coworker install Solaris 10 in a VM.  The last time I used Solaris in any meaningful was was Solaris 7, I think.  But I could be such a geek that I am mixing that up with the arena planet in Battletech…  but back to the point.  I never thought I would see the several things that I saw.  I was very surprised at how well this went given relative ignorance of a Solaris installation.

The most noteworthy surprise I believe is in actually seeing Solaris run on non-SPARC architecture.  That’s been how long it has been since I’ve sat at a SPARCstation and been meaningful with Solaris.  I did a Documentum install/config on Solaris last year, but that is different — if you’ve installed it on one ‘nix, you’ve installed it on them all (except for Solaris’ damn penchant for useless, superfluous directory levels…).

The second biggest surprise is… free.  I’m trying to figure out the catch here.  We downloaded the ISO, installed it, and never was there something that said in so few words “you must pay for this or enter a license key”.  It just worked.

The biggest challenge was in getting the partitions setup the way we needed them.  Of key importance was getting the swap partitions setup so that Oracle (10) would run at all.  I will draw your attention to setting up the swap space.  That was fun.  But that was fun only after I figured out how to partition on Solaris.  It works nothing like Windows or Linux, and I am glad I remembered my acronyms so that c0d0p0s0 made sense to me.   Unlike anything else I’ve done, the partition is not what I expected and the slices are what I expected the partitions to be.  A little trial-and-error helped me figure that one out pretty easily — and a couple of web searches, albeit none very informative about what I was seeing.  Many kudos here to SUN (Oracle) for having such a clear and clean installer.

The swap space was perhaps the strangest thing I have setup.  We made the primary filesystems UFS, and when it came to the swap space, all you do is call it “swap” … just like calling root “/” and any of the other mount points by their appropriate path.  In the end, the three swap spaces we defined (to get 3GB) just simply existed and were implicitly understood to be of fstype “swap”.  Huh…

Finally, I have to say I like what they did with Gnome… at least, I think it is Gnome.  They molded it into a very familiar paradigm (more like KDE3 than typical Gnome), and it was quite easy to navigate.  It still had some of that 1980s feel to it, but how could it not.  It would be a shame if it were to look too bleeding-edge.  {har har}

I will be interested to see the final results.  I have installed Oracle on UNIX before, but never on Solaris.  This is all to facilitate a client in upgrading a system and move it from either “Oracle on SunOS4” (ACK!) or “Oracle on Solaris 2” over to MS SQL Server on Windows Server 2008.  Yes, you read that right — I don’t understand, but I won’t bait for debait {har har}.

Overall, this makes today more interesting than other days, and definitely worth of a technology note.  Solaris in a VM.  Amazing.  I love it.  There is even a VirtualBox appliance that can be downloaded.  I am tempted to do it just for the fun of it because I still get the yearning to tinker away at Solaris from time to time.  It was the ‘nix I cut my teeth on, and it will always have a place in my geeky little heart.


Pseudo Remote Desktop with SSH

This is so simple.  I wish it was this easy on Windows.  And, I wish windows had a clean way to do something like SSH anyway.

local$ xhost+
local$ ssh -X machine-name
remote$ run any gui program

And so, right there, I can run a program on the remote machine and display it back to my local machine.  This is nothing new.   I was doing this 10+ years ago on Solaris systems, albeit done with rsh which is a very insecure way of doing it.

As a note, the above is also insecure.  I should be granting access to specific machines and not using “xhost+” which opens display back from any machine.  But since this is my network, I feel pretty safe.


UNIX find grep

I am always hunting for this, and so I’m going to put it here so I can remember!

This command is great:  find . -exec grep -l hello {} \;

Essentially, it does this:

  1. find all files starting in the current directory (“.”)
  2. for each file found execute “grep -l hello”
  3. grep -l will print file names — with full relative path — that match the grep criteria

This is great, as a programmer, to try and figure out which file had a certain word (or phrase, or Regular Expression pattern) in it.

There are also a couple of other pages with a large set of references here and here.