Kessel Run

I found this trick from a buddy the other day. Check out his blog here

Everyone who uses Unix/Linux uses screen, if not see the MAN page here. I never knew it could open up a screen on a serial device! Turns out it can.

Here is an example for Apple OSX from my friend’s blog .

screen /dev/tty.usbserial 9600

The only thing that would change for Linux/Unix is the device name so just change the device to /dev/ttyUSB0 or /dev/ttyS0.

Just as an FYI comments are disabled, but please feel free to send me an email at any time and I will post an update to the entry itself answering your question. Contact me at {paul{{atxz}kesselrun{{[dot}}{{com]}}

For most of Solaris 10’s short life it has been extremely complex to limit cpu and memory for a zone. Lucky for us the 8/07 update makes things a whole lot easier. I will touch on an easy way to limit disk usage as well.

I am assuming you have a Solaris 10 machine spun up…and have a generic understanding of it’s inner workings.

The below describes a whole-root zone i.e. everything is copied over to the zone. (the create -b is very important)

First off we want to create the config for our zone :

[root@server /]# zonecfg -z testzone
testzone: No such zone configured
Use 'create' to begin configuring a new zone.
zonecfg:testzone> create -b
zonecfg:testzone>

Your config file has been created, it is now time to configure your zone.

zonecfg:testzone> set zonepath=/zones/testzone
zonecfg:testzone> set autoboot=true
zonecfg:testzone> add net
zonecfg:testzone:net> set address=10.10.1.5
zonecfg:testzone:net> set physical=hme0
zonecfg:testzone:net> end
zonecfg:testzone:net> info

That sets up the bare basics of a zone now. The “info” command will list your current setup.

Now things get messy.

CPU limiting (8/07 only!). The process below sets the zone to use only 1 CPU.

zonecfg:testzone> add dedicated-cpu
zonecfg:testzone:dedicated-cpu> set ncpus=1
zonecfg:testzone:dedicated-cpu> end

RAM limiting (8/07 only!). This will cap ram and swap for the zone.

zonecfg:testzone> add capped-memory
zonecfg:testzone:capped-memory> set physical=1g
zonecfg:testzone:capped-memory> set swap=1g
zonecfg:testzone:capped-memory> end


Ok, your ready to go time to commit and exit

zonecfg:testzone> verify
zonecfg:testzone> commit
zonecfg:testzone> exit

You now have a config file in /etc/zones/testzone.xml

Check your running zones.

[root@server /]# zoneadm list -vc
ID NAME STATUS PATH BRAND IP
0 global running / native shared
- testzone installed /zones/testzone native shared


Install your zone…this takes some time.

[root@server /]# zoneadm -z testzone install


Boot up your new zone!

[root@server /]# zoneadm -z testzone boot

Log into your zone’s “console”

[root@server /]# zlogin -C -e [ testzone


Now I glossed over the disk limitations, there are a couple of techniques but I settled on using a separate disk for zones and using slicing (partition) the disk into pieces. You can also do this via filesystem quotas and just letting it fill it’s current mount point.

If I run into any tricks or tips I will post a follow up to this guide. I might try and get running Linux in a zone, however since I use all Sparc64 hardware this seems rather pointless due to the fact branded zones are said to only work in Solaris 10 x86 (I could be wrong…something to test at least.)

Here are a couple of the links I used, this process is not well documented especially with the 8/07 release.

Sun’s Official zone/virtulization guide.
Blastwave’s guide to zones.

I finally got around to setting everything up in my new lab.

Here is a bit of the gear.

Network Gear:

Cisco 2924 24port switch w/ Fiber
Linksys WRT54G w/ dd-wrt
HP 12port Hub (for sniffing and WAN)
Netgear 5port Gigabit switch
Sonicwall Tele3 sp Firewall

Server/Computer Gear:

Parsec : AMD64×2 2gb RAM 2×500gb SATA Linux
Falcon : AMD64 1gb RAM 1×400gb SATA Linux/Windows
Osiris : Toshiba Sat. P4M 512mb RAM 1×40gb IDE Linux/Windows
Ra : Sun Netra X1 1gb RAM 2×40gb IDE Solaris 10
Chewie : IBM T30 P4M 512MB RAM 1×60gb IDE Linux
Carbonite : IBM X60 Core2duo 1gb RAM 1×80gb SATA Linux
Maw : Sun Netra X1 256mb RAM 2×40gb IDE Openbsd
Dreadnaught : P4 XEON x2 6gb RAM 3×73gb SCSI Linux
unnamed : P4 XEON 2gb RAM 2×36gb SCSI Linux

Cheers!

In case anyone has ever wanted to try Absinthe, it is now legally in the states under the names Lucid (France) and Kubler (Switzerland).

For only “legally” drinking for about 1 year, I think I am not doing so bad….

It took me some time to get things back up and running (pull backups, move servers, etc.) But ck3k.org is backup! I doubt anyone noticed or cared, but this makes it look like I keep this blog up to date

-Paul

“On Tuesday, November 13, and Thursday, November 15, the two-part Star Trek Remastered version of “The Menagerie” will beam onto the big screen in a special engagement with selected theatres.” - startrek.com

I am going, that is all there is to it. I am also going with a fellow trekkie WildBill , so should be fun. I am having trouble deciding….what shirt should I wear? See below for options….

I am de-racking my server for ck3k.org (friend is moving), so it will redirect here for the moment. I should have it back up by the end of the day. If anyone cares or needs something from it, let me know.

-Paul

In reply to my last post, I went ahead and got the VW GTI, the thing drives like a dream. I will have to upload pics soon! I am also waiting for the DMV to deny my custom plate request of “J3D1.”….This is not the GTI you are looking for….

I have been driving around my 03 Eclipse GS for about a year now, and it is starting to show the abuse that I put it though during my commute. I am looking to upgrade. Here is the list of the cars I think rock, keeping in mind I am lame and can’t drive manual…yet

My budget is from 18,000 to about 26,000 dollars.

2005-2006 Subaru WRX
2006-2007 VW GTI
2001-2002 BMW Z3
2007 Mini Cooper S
2004-200* BMW 325ci
2004-200* Infinti G35
2004-200* Nissan 350z

I like German cars, but I am worried about the costs of keeping them maintained. If they where cheap I would have alread gotten a 3-series BMW. If anyone has any ideas, let me know!

This is a technique for failing over a pair of nics on a Solaris 10 machine. I am told this will only work with Sun nics that come with Sun hardware, but I haven’t tested it.

You need to be root for all of the following, I am also assuming your going to be able to reboot the machine once this setup is in place.

First make sure you have your hostname in your /etc/hosts file


[ck3k@ra ~]$ cat /etc/hosts
#
# Internet host table
#
127.0.0.1 localhost
10.10.1.112 ra loghost


Ok, so in your /etc/hosts ra is your hostname. Now we need to edit/create two files. NOTE the .dmfe* will be the interface name, for example on a SunFire T2000 it is hostname.ipge*.

First setup your main nic’s hostname file (This should already exist in your install)


[ck3k@ra ~]$ cat /etc/hostname.dmfe0
ra netmask + broadcast + group failover up


So you see your hostname and it’s ip info, but now it has a group called “failover.” This creates a group of nics that will failover for one another.

Now create the hostname file for your second nic.


[ck3k@ra ~]$ cat /etc/hostname.dmfe1
group failover up


This just adds the nic to the group for failover. Now, reboot your system and the settings will take affect on boot. Keep in mind this will work for virtual interfaces as well, so all of your Zones will also failover.

Once your system reboots, the best way to test is to be on the LOM via serial and to pull one of your two nics out of your switch.

Here some of the output of the nic going down :


Aug 16 17:54:28 ra dmfe: [ID 801593 kern.notice] NOTICE: dmfe0: PHY 1 link down
Aug 16 17:54:28 ra in.mpathd[135]: [ID 215189 daemon.error] The link has gone down on dmfe0
Aug 16 17:54:28 ra in.mpathd[135]: [ID 594170 daemon.error] NIC failure detected on dmfe0 of group failover
Aug 16 17:54:28 ra in.mpathd[135]: [ID 832587 daemon.error] Successfully failed over from NIC dmfe0 to NIC dmfe1
Aug 16 17:55:05 ra dmfe: [ID 801593 kern.notice] NOTICE: dmfe0: PHY 1 link up 100 Mbps Full-Duplex


Here it is coming back up :


Aug 16 17:55:05 ra dmfe: [ID 801593 kern.notice] NOTICE: dmfe0: PHY 1 link up 100 Mbps Full-Duplex
Aug 16 17:55:05 ra in.mpathd[135]: [ID 820239 daemon.error] The link has come up on dmfe0
Aug 16 17:55:05 ra in.mpathd[135]: [ID 299542 daemon.error] NIC repair detected on dmfe0 of group failover
Aug 16 17:55:05 ra in.mpathd[135]: [ID 620804 daemon.error] Successfully failed back to NIC dmfe0


You now have a redundant Solaris machine that can resist some layer one failures.