My Teneo!!!

Is Adobe's Creative Cloud a Good Thing

2013-05-16T22:35:55Z

Adobe's Creative Suite (Photoshop, Illustrator, InDesign, Flash, etc) is probably the most pirated software in the world (my opinion). In many industries, it's considered the de-facto standard for today's graphics, video and media productivity tools. So, obviously, Adobe had to do something to get their piracy under control, and re-establish some of that missing monetary value.

The chosen route? Monthly subscriptions to their software. This has the benefit of only paying for what you need. You pay a monthly amount for each month you want it. New upgrades? No problem. Just get it as needed. It's not hosted on the internet, so you just download and install. Plus, the costs are usually lower over the lifespan of the product that you use.

Let's assume you want the latest Photoshop. One website listed it as $728. Also, within 3 years, you probably would have purchased an upgrade for $200. But, Creative Cloud would sell it to you for $20 per month. It would take 46 months of usage to equal that price, and you would get some cloud storage and upgrades out of it.

Sounds great. Why wouldn't you? Well, many smaller businesses (SMB market) that are privately owned are usually measured by their EBIDTA as to the health and profit of the company. Those businesses would rather spend Capex over Opex, since Capex is depreciated over time. And software licenses can be applied to Capital Purchases, since it gives value to the company.

However, now that Adobe has gone to a subscription-based model, you can no longer apply this to CapEx. It must be opex, and it hits your expense budget immediately. Now, for $20/month, you can probably afford it. Unless your company has a large adobe team.

Overall, I can see this being a benefit. But, I can also see this being a nightmare of IT centralization of license usage and purchasing, along with budgets. Especially since this came out of no-where, and in the middle of the year where many companies are only half-way through their already approved budgets.

It will be interesting to see how it all plays out.

**UPDATE** - Check out Terri White's 5 Myths to Adobe's Creative Cloud: http://terrywhite.com/5-myths-about-adobe-creative-cloud/

Citrix Netscaler, Synergy, and LPS Integration

2013-05-14T18:10:53Z

Last year, I decided it was time to start working on virtualization at the desktop. The reasons were numerous and plentiful, and beyond the scope of this writing. Needless to say, it is a good decision.

I decided on the Netscaler 8200 for it’s ability to scale, and for the Citrix Access Gateway features. Man, this device is quite powerful. Enough so, for me to question why I’m running an F5 for my load balancing. Sure, the F5 is really powerful and loaded with options. But, I don’t use those options. I just use regular load balancing algorithms.

After reviewing the costs, the maintenance on 2 Netscaler’s for High Availability hits roughly about the same cost as one F5. Wait, I can get high-availability, SSL VPN, Citrix App Delivery, AND still load balance without increasing my OpEx? Sign me up! So, do I really need an F5? Not really. The F5 is a fantastic box, and just works, but overall, I’m going for costs and usability.

So, since then, I’ve been deploying Citrix XenApp along with the mobile receivers and the HP t410/t510 clients. However, there have been quite a few hiccups on the road. Not necessarily from an implementation perspective, just common issues. And being illiterate in Citrix, resolutions have been incredibly slow. So, I reached out to a local provider for some Citrix training. Enter LPS Integration.

Now LPS Integration is a platinum partner in Citrix, and is local to me. They've got some really smart people, and very easy to talk to. After inquiring about some training, they decided I could get a better understanding of what they can do for me, if I go to Citrix’s annual tech conference, called Synergy. So, they are sponsoring me the conference pass.

Now, I’m super stoked. I have signed up for as many sessions that are still open, and will blog about as many as I can. Thanks LPS!

If anyone is going, and want to hook up, send me a message! @Neelixx. You can also watch (or filter) my tweets using #CitrixSynergy

Flickr VS Instagram

2013-04-30T22:37:06Z

Everyone remembers Flickr, right? It used to be the primo photo sharing service out there. But, it was complex, slow to learn, and even slower to share. It really had a "Photographer" feel to it. In other words, if you weren't in the profession, Flickr didn't feel right.

When initially started, Instagram took off like wildfire? Why? It was easy. You take a picture, put a filter on it, and post it. Simple. And people love simple. So, why change?

If memory services, back in December of 2012, Instagram worded their new terms of services in a way that would allow them to use your photos for advertising, and possibly even sales. Would they? Who knows... but the legal jargon allowed them to. So, lots of people went into an uproar about it and got Instagram to put things back.

In my opinion, the damage was done. Then, it got me thinking about copyright, and how I loved the idea that Flickr gave me copyright options (Creative Commons). But, why would I go back to Flickr? It was not exactly easy. But, oh, so it is. They learned from Instagram and made their own mobile app just as easy (but using "advanced" can give you the more control). Click, Filter, Publish. Done! Need to share it to Facebook or Twitter or <insert social network service>? Yep, you can do that too. You can even embed the images into a website:

But, if you want to get more detailed, in the same post, you can filter, tag, assign sets, to multiple images at the same time. You can also browse your "contacts" images like Instagram.

While I am still learning and getting a "feel" for Flickr, I really like the changes, and have already started moving my images over. Will you? Tell me your thoughts.

Add a drive to an existing Linux LVM

2013-04-12T21:54:19Z

Had a problem where I ran out of disk space, and needed to increase it on a Linux server. Now, I'm not a Linux engineer, but I have used it for some time. As far as I know, there are 2 ways to extend a drive that is built using LVM (Logical Volume Manager).

Add an additional drive, add it to LVM, and expand the data to use the new drive. (physical and VMware)
Expand the drive in VMware, then expand the LVM (VMware only)

I'm going to discuss option 1, since it can be used for both physical and virtual.

First off, let's see what we have to work with, shall we? Let's do a "df -h" to see our current disk usage:

[root@www2 ~]# df -h

`Filesystem`	`Size`	`Used`	`Avail`	`Use%`	`Mounted on`
`/dev/mapper/VolGroup00-LogVol00`	`15G`	`14G`	`0`	`100%`	`/`
`/dev/sda1`	`99M`	`13M`	`82M`	`14%`	`/boot`
`tmpfs`	`1.7G`	`0`	`1.7G`	`0%`	`/dev/shm`

Now, do a scan of our volume group:

[root@www2 ~]# vgscan

Reading all physical volumes. This may take a while...

Found volume group "VolGroup00" using metadata type lvm2

Let's see what our LVM Disks are:

[root@www2 ~]# lvmdiskscan

/dev/ramdisk [ 16.00 MB]

/dev/root [ 15.00 GB]

/dev/ram [ 16.00 MB]

/dev/sda1 [ 101.94 MB]

/dev/VolGroup00/LogVol01 [ 768.00 MB]

/dev/ram2 [ 16.00 MB]

/dev/sda2 [ 7.90 GB] LVM physical volume

/dev/ram3 [ 16.00 MB]

/dev/ram4 [ 16.00 MB]

/dev/ram5 [ 16.00 MB]

/dev/ram6 [ 16.00 MB]

/dev/ram7 [ 16.00 MB]

/dev/ram8 [ 16.00 MB]

/dev/ram9 [ 16.00 MB]

/dev/ram10 [ 16.00 MB]

/dev/ram11 [ 16.00 MB]

/dev/ram12 [ 16.00 MB]

/dev/ram13 [ 16.00 MB]

/dev/ram14 [ 16.00 MB]

/dev/ram15 [ 16.00 MB]

/dev/sdb [ 8.00 GB] LVM physical volume

3 disks

16 partitions

1 LVM physical volume whole disk

1 LVM physical volume

[root@www2 ~]#

And finally, our hard drives:

[root@www2 ~]# fdisk -l

Disk /dev/sda: 8589 MB, 8589934592 bytes

255 heads, 63 sectors/track, 1044 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/sda1 * 1 13 104391 83 Linux

/dev/sda2 14 1044 8281507+ 8e Linux LVM

Disk /dev/sdb: 8589 MB, 8589934592 bytes

255 heads, 63 sectors/track, 1044 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdb doesn't contain a valid partition table

[root@www2 ~]#

At this time, I insert the hard drive. This is done by adding another drive to the physical machine, or adding a drive using VMware. Either way, add it, and reboot.

Now, when you do a fdisk -l, you will see a listing of drives, and my new drive (which is /dev/sdc)

[root@www2 ~]# fdisk -l

Disk /dev/sda: 8589 MB, 8589934592 bytes

255 heads, 63 sectors/track, 1044 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Device Boot Start End Blocks Id System

/dev/sda1 * 1 13 104391 83 Linux

/dev/sda2 14 1044 8281507+ 8e Linux LVM

Disk /dev/sdb: 8589 MB, 8589934592 bytes

255 heads, 63 sectors/track, 1044 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdb doesn't contain a valid partition table

Disk /dev/sdc: 32.2 GB, 32212254720 bytes

255 heads, 63 sectors/track, 3916 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Disk /dev/sdc doesn't contain a valid partition table

Now, let's create a partition

[root@www2 ~]# fdisk /dev/sdc

Device contains neither a valid DOS partition table, nor Sun, SGI or OSF disklabel

Building a new DOS disklabel. Changes will remain in memory only,

until you decide to write them. After that, of course, the previous

content won't be recoverable.

The number of cylinders for this disk is set to 3916.

There is nothing wrong with that, but this is larger than 1024,

and could in certain setups cause problems with:

1) software that runs at boot time (e.g., old versions of LILO)

2) booting and partitioning software from other OSs

(e.g., DOS FDISK, OS/2 FDISK)

Warning: invalid flag 0x0000 of partition table 4 will be corrected by w(rite)

Command (m for help): n

Command action

e extended

p primary partition (1-4)

>p

Partition number (1-4): 1

First cylinder (1-3916, default 1): 1

Using default value 1

Last cylinder or +size or +sizeM or +sizeK (1-3916, default 3916): 3916

Using default value 3916

Command (m for help): t

Selected partition 1

Hex code (type L to list codes): 8e

Changed system type of partition 1 to 8e (Linux LVM)

Command (m for help): w

The partition table has been altered!

Calling ioctl() to re-read partition table.

Syncing disks.

[root@www2 ~]#

Let's look at our disks, now that we have added a new drive and partitioned it. Did our LVM Disks change?

[root@www2 ~]# lvmdiskscan

/dev/ramdisk [ 16.00 MB]

/dev/root [ 15.00 GB]

/dev/ram [ 16.00 MB]

/dev/sda1 [ 101.94 MB]

/dev/VolGroup00/LogVol01 [ 768.00 MB]

/dev/ram2 [ 16.00 MB]

/dev/sda2 [ 7.90 GB] LVM physical volume

/dev/ram3 [ 16.00 MB]

/dev/ram4 [ 16.00 MB]

/dev/ram5 [ 16.00 MB]

/dev/ram6 [ 16.00 MB]

/dev/ram7 [ 16.00 MB]

/dev/ram8 [ 16.00 MB]

/dev/ram9 [ 16.00 MB]

/dev/ram10 [ 16.00 MB]

/dev/ram11 [ 16.00 MB]

/dev/ram12 [ 16.00 MB]

/dev/ram13 [ 16.00 MB]

/dev/ram14 [ 16.00 MB]

/dev/ram15 [ 16.00 MB]

/dev/sdb [ 8.00 GB] LVM physical volume

/dev/sdc1 [ 30.00 GB]

3 disks

17 partitions

1 LVM physical volume whole disk

1 LVM physical volume

Now, let's create a physical volume for the new partition of the new drive

[root@www2 ~]# pvcreate /dev/sdc1

Physical volume "/dev/sdc1" successfully created

Remind me again which Volume Groups we have? We need to extend it to cover the new drive

[root@www2 ~]# vgscan

Reading all physical volumes. This may take a while...

Found volume group "VolGroup00" using metadata type lvm2

Oh yeah, that's right. VolGroup00. Lets extend the volume to include the new partitioned drive.

[root@www2 ~]# vgextend VolGroup00 /dev/sdc1

Volume group "VolGroup00" successfully extended

Okay. The Volume Group has been extended. Now, we need to extend the Logical Volume. Where is this LogVol? I need to know what to expand.

[root@www2 ~]# lvdisplay

--- Logical volume ---

LV Name /dev/VolGroup00/LogVol00

VG Name VolGroup00

LV UUID O9VkzE-BOec-nvse-7yPD-ebuR-2BTK-2DYqWk

LV Write Access read/write

LV Status available

# open 1

LV Size 15.00 GB

Current LE 1440

Segments 3

Allocation inherit

Read ahead sectors auto

- currently set to 256

Block device 253:0

Here, I am adding 30GB ( -L+30G) to the logical volume. I know this, because I added a 30GB drive. There are other ways to extend it, but I'm just using size.

[root@www2 ~]# lvextend -L+30G /dev/VolGroup00/LogVol00

Extending logical volume LogVol00 to 45.00 GB

Logical volume LogVol00 successfully resized

Now that the Logical Volume is resized, we need to expand the current file system.

[root@www2 ~]# resize2fs /dev/VolGroup00/LogVol00

resize2fs 1.39 (29-May-2006)

Filesystem at /dev/VolGroup00/LogVol00 is mounted on /; on-line resizing required

Performing an on-line resize of /dev/VolGroup00/LogVol00 to 11796480 (4k) blocks.

The filesystem on /dev/VolGroup00/LogVol00 is now 11796480 blocks long.

Done!!

UCCX Scripting - Iterating Different Recordings While on Hold

2013-01-25T04:12:29Z

When scripting UCCX call center queues, it's common to put a caller on hold for a defined period of time, then do something with them. Then, put them on hold again. This could be to hear a "Thanks for holding. We'll get with you shortly" every 2-3 minutes with hold music in between.

Basic

For example, here, you will see a basic call loop. A call gets queued if the resource (call center queue) is not available. I created a variable called "CallQueueHoldDelay". This is set to 90. Which means, the call is put on hold (hears hold music), waits 90 seconds, then goes off hold. I play a "Thanks for holding message", then, goes back to "QueueLoop", and does it all over again.

But, for the script I'm working on now, I have different prompts I want caller's to hear. Maybe different accouncements? Different sales going on? Any alerts caller's should be aware of? Well, you can play these back to back, but that's information overload. We want our callers to be calm and informed.

So, how do we space them out? Welcome to the math operator called Modulus (or Modulo). Using the %, you basically get the remainder from the result of division. For example, If I divide 2 / 3... I get 1 with a remainder of 1, right? If I divide 10 / 4, I get 2 with a remainder of 2 ( 4X2=8 with 2 remaining).

In other words, the modulus operator defines how many "options" we can have before going back to the start. Let's do some examples:

Even / Odd

This basically toggles between one or the other. 2 different options. So, let's say you have 2 announcements. You want someone to hear announcement one, hold, then hear announcement two. Looking at my script, you can see I am incrementing "CountQueueLoop" by 1 each time. So, assuming CountQueueLoop starts at 0, when I go on hold, I SWITCH based on the value of CountQueueLoop % 2 (which is 0%2 which is 0). That will hit "Iteration 1" (since the value is 0) and run those commands. The next time I come back, CountQueueLoop is now 1 (which is 1%2 which is 1). That will hit Iteration 2. The 3rd time, CountQueueLoop is 2 (which is 2%2 which is 0). So, the results are only 0 and 1.

3 options

Below, you can see the same thing, only I use 3 options ( CountQueueLoop % 3). Because I am dividing by 3, I have up to 2 remainders giving me 3 options. If I use CountQueueLoop % 5, how many options can I have? 5. That's right! See, you are a natural.

In my case, I am toggling between 2 announcements. However, I don't want them back to back, so I'm using 3 iterations so that the third iteration will just be more hold to space them out a bit more.

Good Luck!

Static Mapping of Serial Devices in Linux

2012-12-16T20:30:43Z

Some time ago, I wrote about how to setup a Linux-based console server. But, there is one problem… if you reboot, your serial devices may have moved. So, to fix that, we will use UDEV to statically map the serial devices to non-changing names.

Plug in your usb converter and run the command ‘dmesg’. Look for where it was mapped to. In my case, it is ttyUSB1

usb 3-2: new full speed USB device using uhci_hcd and address 5
usb 3-2: configuration #1 chosen from 1 choice
pl2303 3-2:1.0: pl2303 converter detected
usb 3-2: pl2303 converter now attached to ttyUSB1

Now, the problem is, I plan on mapping ttyUSB1 to be connected to Switch3, but the next time I reboot, this converter could be mapped to ttyUSB0 (which may go to SW1 in my configs), and everything would be befungled.

So, we need to map it to a static name that will always point to the right switch. I will call the devices ttyCon1 thru ttyCon6. Here is how, and we will use USB1 to start.

First, we need to find some attributes to use to uniquely identify this specific adapter. Since we have multiple adapters from the same Manufacturer, it’s best we use the SERIAL number assigned to these devices. Let’s get it from the command ‘lsusb -v’, which will list all USB devices verbosely.

We only care about the “Prolific” adapters, and the device descriptor (since we are identifying at the ‘device’ level)

Bus 002 Device 003: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x067b Prolific Technology, Inc.
  idProduct          0x2303 PL2303 Serial Port
  bcdDevice            4.00
  iManufacturer           1 Prolific Technology Inc. 
  iProduct                2 USB-Serial Controller D
  iSerial                 0 
  bNumConfigurations      1

Bus 003 Device 005: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x067b Prolific Technology, Inc.
  idProduct          0x2303 PL2303 Serial Port
  bcdDevice            4.00
  iManufacturer           1 Prolific Technology Inc. 
  iProduct                2 USB-Serial Controller D
  iSerial                 0 
  bNumConfigurations      1

Bus 004 Device 002: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x067b Prolific Technology, Inc.
  idProduct          0x2303 PL2303 Serial Port
  bcdDevice            4.00
  iManufacturer           1 Prolific Technology Inc. 
  iProduct                2 USB-Serial Controller D
  iSerial                 0 
  bNumConfigurations      1

Bus 004 Device 003: ID 067b:2303 Prolific Technology, Inc. PL2303 Serial Port
Device Descriptor:
  bLength                18
  bDescriptorType         1
  bcdUSB               1.10
  bDeviceClass            0 (Defined at Interface level)
  bDeviceSubClass         0 
  bDeviceProtocol         0 
  bMaxPacketSize0        64
  idVendor           0x067b Prolific Technology, Inc.
  idProduct          0x2303 PL2303 Serial Port
  bcdDevice            4.00
  iManufacturer           1 Prolific Technology Inc. 
  iProduct                2 USB-Serial Controller D
  iSerial                 0 
  bNumConfigurations      1

Now, here is where you would look at the “iSerial” to uniquely identify each adapter. But, you can see, I have a problem. Looks like the everything is the same. Even the serial number (really cheap devices??). We need to use something UNIQUE to identify each adapter. I guess the only way is to use the ID or port that it’s plugged into.

It would probably be best to use the Vendor, Product, and Serial to identify your adapters.

We will use udevinfo for that. Use this for each ttyUSB# that you have attached to your system.

[root@sys1 ~]# udevinfo -a -p /sys/bus/usb-serial/devices/ttyUSB1

looking at device '/devices/pci0000:00/0000:00:1d.1/usb3/3-2/3-2:1.0/ttyUSB1':
  KERNEL=="ttyUSB1"
  SUBSYSTEM=="usb-serial"

looking at parent device '/devices/pci0000:00/0000:00:1d.1/usb3/3-2/3-2:1.0':
  ID=="3-2:1.0"
  BUS=="usb"
  DRIVER=="pl2303"
  SYSFS{modalias}=="usb:v067Bp2303d0400dc00dsc00dp00icFFisc00ip00"
  SYSFS{bInterfaceProtocol}=="00"
  SYSFS{bInterfaceSubClass}=="00"
  SYSFS{bInterfaceClass}=="ff"
  SYSFS{bNumEndpoints}=="03"
  SYSFS{bAlternateSetting}==" 0"
  SYSFS{bInterfaceNumber}=="00"

So, I’m going to use the ID==”3-2:1.0”. I could be wrong, but I read this as “Hub #3, port #2, device #1. Which means, I’m mapping the physical port, not the adapter. Of course, if I ever MOVE this adapter or hub, everything changes. This is why it’s best to use the Serial rather than the port. Since this is only for my lab, I’m ok with that.

UDEV Rules are put in the directory /etc/udev/rules.d/, and are processed alphabetically (and must end in .rules). So, I’m going to call mine 45-swcon.rules.

BUS=="usb", ID=="3-2:1.0", SYMLINK+="ttyCon2"

Doing this, maps the adapter in 3-2:1.0 to be /dev/ttyCon2. So, if my USB adapter changes to ttyUSB102, it will still be mapped to ttyCon2, and THAT is what I use in my configs.

I can now do this for my remaining adapters. My file /etc/udev/rules.d/45-swcon.rules file now looks like this:

BUS=="usb", ID=="2-2:1.0", SYMLINK+="ttyCon1"
BUS=="usb", ID=="3-2:1.0", SYMLINK+="ttyCon2"
BUS=="usb", ID=="4-1:1.0", SYMLINK+="ttyCon3"
BUS=="usb", ID=="4-2:1.0", SYMLINK+="ttyCon4"

Now, this maps the adapter based on port. Obviously, this isn’t very flexible, but it’s all I can do, since these are VERY cheap adapters. You get what you pay for, right? If your adapters have serial numbers, use that instead. Maybe something like:

BUS=="usb", SERIAL=="0a123bcd4e1cba", SYMLINK+="ttyCon1"

That way, no matter where you plug this bad boy in, it will always be mapped to ttyCon1.

Good luck!

Create your own Serial Terminal Server

2012-12-16T19:41:27Z

Have you ever had the need to have a terminal server to access your serial devices? Maybe it's a Cisco switch, a Juniper router, maybe some HP equipment. If it is accessed through a serial connection, you can make your own Linux-based serial terminal server, without having to purchase OpenGear or Perle, or building a Cisco terminal access server from an old 2500 octal cable. All you need is a Linux system, and some serial ports.

Now-a-days, 9-pin serial ports are hard to come by, even on today's desktop systems. Even if you are lucky to find an older system, chances are, you'll have at most, 2 ports. Here is where USB-to-Serial converters come into play. Buy some, and plug them in. Linux will recognize them, and create the serial devices for them. Don't have enough USB ports? Buy a 10-port USB hub! :)

Materials:

TRENDnet TU-S9 USB-to-Serial adapter - Ebay - $9.00 new
CentOS 5.8 desktop system

Step 1: Install SER2NET in CentOS/RHEL. This is a proxy service that allows your network telnet sessions to be redirected to locally attached serial devices

Install Extra Packages for Enterprise Linux (EPEL)

For CentOS/RHEL 5.x

rpm -Uvh http://dl.fedoraproject.org/pub/epel/5/i386/epel-release-5-4.noarch.rpm

For CentOS/RHEL 6.x

rpm -Uvh http://download.fedoraproject.org/pub/epel/6/i386/epel-release-6-5.noarch.rpm

Install ser2net
1. ```
yum install ser2net
```

Step 2: Identify your serial adapters in CentOS

This seems to be a bit of a frustration to some users. Finding *which* USB port goes to which device. There are a couple of ways to do this, but the easiest is to plug in one-at-a-time and identify what the Operating System identified it to be. As you plug in your usb device, within a second or two, you can run the following command:

  dmesg | grep tty

This will pull up all the messages associated with the serial ports. The last line would be the one you just added:

     [root@sys1 ~]# dmesg | grep tty 
                serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A 
                00:07: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A 
                usb 4-1: pl2303 converter now attached to ttyUSB0 
                usb 4-2: pl2303 converter now attached to ttyUSB1 
                usb 3-1: pl2303 converter now attached to ttyUSB2 
                usb 2-2: pl2303 converter now attached to ttyUSB3 
                usb 3-2: pl2303 converter now attached to ttyUSB4

In this case, the last device I added was attached to "ttyUSB4", which would make the virtual device /dev/ttyUSB4 used in the configuration steps below.

Step 3: Configure ser2net

The configuration files are located in /etc/ser2net.conf. Here is a snippet of my configuration:

NOTE: The "##" lines define comments, and are not needed

## This defines a banner called "cisco-sw1". Places some line breaks, tells me a message, and the device I'm connected to
BANNER:cisco-sw1:\r\n\r\n\r\nYou are connected to C-SW1 on device \d \r\n\r\n 
BANNER:cisco-sw2:\r\n\r\n\r\nYou are connected to C-SW2 on device \d \r\n\r\n 
BANNER:cisco-sw3:\r\n\r\n\r\nYou are connected to C-SW3 on device \d \r\n\r\n 
BANNER:juni-sw1:\r\n\r\n\r\nYou are connected to J-SW1 on device \d \r\n\r\n


## This creates a telnet port 4001, with no timeout, to go to /dev/ttyS0 with a speed of 9600 baud and use the banner named "cisco-sw1"
4001:telnet:0:/dev/ttyS0:9600 cisco-sw1 
4002:telnet:0:/dev/ttyUSB0:9600 cisco-sw2
4003:telnet:0:/dev/ttyUSB1:9600 cisco-sw3 
4004:telnet:0:/dev/ttyUSB2:9600 juni-sw1

The BANNER is optional, but it tells me what I've connected to, once I establish my telnet session. Be sure to restart ser2net service after changing your configuration file! "service ser2net restart"

So, now, I just: "telnet <ip of your server> 4001" to gain access to the console port of that switch/router. See below for tweaks.

Tweaks:

So, you want to use this in your production environment? Using Telnet may not be the best way, unencrypted passwords/text and all. Just use SSH into your system, and telnet to the local host port:

     telnet localhost 4001

Don't want to remember all those ports? Write up a quick alias:

     alias SW1="telnet localhost 4001"
        alias SW2="telnet localhost 4002"

If you don't remember what you called an alias, because you are a slacker and haven't console'd in awhile, just use the "alias" command by itself, and it will list all the entries for you.

Can't authenticate Cisco to HP iMC Tacacs

2012-12-04T15:29:22Z

I love TACACS+. I know there are those of you who prefer RADIUS, but there are a few reasons why I love Tacacs

It's tried and true. Stable
It's TCP not UDP (c'mon, you knew I'd use this...)
It's easiest to configure Group -> Command mapping

So, when HP iMC 5.1 came out with TAM (Tacacs Authentication Module), I just HAD to try it out! So, I grabbed a lab switch, downloaded the trial version of TAM, and started hacking away. It wasn't long before I was stuck. No matter what I did, I just couldn't authenticate to iMC.

I checked the host firewall. TCP49 should be open.
I made sure the device is listed in the "Device List" of iMC TAM Manager.
I made sure tam.exe is binding to the right IP Address, using "netstat -ban".
I made sure I have created a device user, and that device user is assigned the right device user group, and that group has the permitting authorization profile.

I checked my firewall, used Wireshark, even threw in some AAA / Tacacs debug commands. Still, nothing.

I still can't login. In order to troubleshoot more, I needed to dig a little deeper. Going to the log file for TAM (C:\Program Files\iMC\tam\log) I found the following entry:

% 2012-12-03 14:55:08 ; [WARNING (2)] ; [3872] ; TAM ; $SYS$ ; (NULL) ; (NULL) ; (NULL) ; Invalid Source IP or port number(from 192.168.1.253:49).

HP iMC apparently does not include ALL IP addresses of a device when matching. More than likely, iMC uses the numerically lowest IP Address as it's primary address to identify and collect on. This is the IP Address that it is expecting to see. No other address is valid. So, even though my device in iMC has the address of (VLAN 10: 10.10.10.1), TAM is denying it, because it is coming from the IP of 192.168.1.253 (which is on VLAN 1).

How to fix it? Manually tell Cisco to use the iMC defined IP Address:

ip tacacs source-interface vlan 10

Now, why is the Cisco using that IP Address in the first place? Well, in my lab, the iMC is on the same subnet as VLAN 1. So, logically, the Cisco will send out the packet from the closest interface.

How I label network cabling

2012-11-29T20:37:02Z

Recently, I was asked what the best way is to label your cable plant within an office/building/campus environment. Of course, my answer was, "it depends".

Scenarios

The first thing I think about is "scalabililty". When you come up with your own "scheme" to label your cables, it needs to expand well. It also needs to handle changes easily, and with the least amount of time to do. For example, lets say you have 1 wiring closet, and you start labeling your cables starting from 1 (or 001, if your OCD requires you to pad your numbers). After 3 months, you have 2 racks full and your cable numbers go up to 300, and had to start a new wiring closet on the other end of the building.

Lessons Learned:

It takes time to relabel each port of the patch panels with the new cable number.
You have to remember what cable number you left off at when terminating new cables.
No easy way to find out which wiring closet a jack goes to.
You spend more time labeling, than actually terminating the cables.

You modify your approach to include office and cube names, to make it easier to identify the location of the cable. You also setup a system, so that each new wiring closet will start with a number. (i.e. Wiring Closet 1 = 1xx, Wiring Closet 2 = 2xx, etc). You also start labeling your server cables at the switch ports, to identify which cable goes to which server / switch.

Lessions Learned:

Each wiring closet is limited to 99 cables. Ooops.
Office changes forced the removal of 3 cubes that were labeled. 1 office is now a conference room. The location labels no longer are accurate.
After installing a Cisco 6509 with 7 48-port blades, you can't even see your labels in the big blur of color. Even the best cable management can't seperate the cables far enough to see the labels you put on.

Conclusion

Here is how I do it. My way may not be the best way, but it has worked for me this long. Here are my stipulations:

Don't label the patch panel. Too many ports and it wastes too much time.
Try and be as generic as possible, while still being accurate. Using "Server1" or "Cube2" is too specific, and can change.
Count Patch Panels from top down. Each Panel that has a "Port 1" gets a count of "1". (Had to add this, because some places had 24-, 48-, and 96-port panels, and they didn't know how to count them).
When labeling cables at servers and switchports, bring the label back 3-6 inches from the end. It will make identifying the cable much easier in dense environments. But stay consistent. It is not very aesthetic to see some cables at 3 inches, and some at 6 inches.
Use a cable labeler, like Dymo or Brady IDXpert. IDXpert is my favorite, because it can print on heat shrink tubing (the best for labeling cables), or wrap-around labels.

Here is how I label.

Getting user-mode after logging in with aaa user at privilege 15

2012-10-15T20:29:53Z

Had an interesting issue today. I enabled AAA Authentication on a used Cisco 3560 switch. I then created a user with privilege 15. But, every time I used either telnet or SSH, I was always getting User Mode. I wanted Priveleged Mode.

Since the highest level of privilege commands is 15, I should be getting full permissions if my user is ALSO set at privilege 15, right? Well, it does, but first there was a tiny configuration that needs to happen. First, let's enable AAA on the device:

big-old-switch(config)# aaa new-model

Now, let's create a user:

big-old-switch (config)# username neteng privilege 15 secret blahblahblah

Now, I have a user account with privilege level 15. All I need to do is telnet to my device, and I'm golden, right? Let's do that:

[root@host ~] telnet 1.2.3.4 

Trying 1.2.3.4... 
Connected to big-old-switch (1.2.3.4). 
Escape character is '^]'. 

User Access Verification 
Username: neteng 
Password: 

big-old-switch>

Wait, what? Why am I not in enable mode? I should have privilege 15, right? Let's check:

big-old-switch>show priv
Current privilege level is 1
big-old-switch>

Now, hold on a sec. Cisco, have you gone stupid? I just created a user account with privilege 15. I know.... because I just logged in as him. What gives?

The Answer:

Well, for starters, I called Cisco stupid. These devices can be *very* temperamental. Once I got off my high horse (and asked on Twitter), I realized I didn't have any authorization statements. The winning statement here?

	
big-old-switch(config)# aaa authorization exec default local

This sets the exec shell level according to AAA, which, in my case is 15. NOW, let's try it again!

[root@host ~] telnet 1.2.3.4 
Trying 1.2.3.4... 

Connected to big-old-switch (1.2.3.4). 
Escape character is '^]'. 

User Access Verification 
Username: neteng 
Password:  

big-old-switch# 
big-old-switch#show priv 
Current privilege level is 15 
big-old-switch#

Eureka! Thanks Twitter! (actually @revolutionwifi and @xanthien). There were others, but these two got it first. I hope this helps someone else. I didn't see any posts, so either I'm the only one who has had this problem, or someone just hasn't written about it yet. Since AAA has been around for a long time, I'm willing to bet, it's just me having an off day. Yeah, that's it. I'm getting more coffee.

The use-case for PoE thin clients

2012-10-11T16:41:31Z

2 months ago, I started a project to move forward with Virtual Desktops at my company. Obviously, one of the components of that is the thin clients. Who to use? Why? The industry seems to be swarming with them. Wyse, HP, NComputing, IGEL, 10zig, etc. So, it becomes a tough call.

A discussion started about someone being at Citrix Synergy 2012 and saw a full PoE display and client! I was floored! Why? To run a 19" display WITH the client AND USB, and under 15 watts? That sounds unheard of. And, to be frank, I didn't believe it until I saw it for myself. Once I saw it, a twitter discussion was not long after that.

As I was talking to an NComputing guy about the sexiness of the HP T410, he then made the comment that no one really made a use-case for this type of product. Really? You can't see it? Well, here are my reasons why this is just too cool for words:

Greater Control - You will never need to ask the user to unplug the client's power. You can do it. No need to ask anymore.
Disaster Recovery - If your building goes dark due to a power outage, your UPS and/or generator for your equipment room will continue to power the devices.
Cheaper Move/Add/Changes - No need to call an electrician to run new electrical conduits. Just drop a Cat5. You are done! Moving offices? Your team can handle it, without the added expense or time for electricians.
Longetivity - By removing the power brick from the device, you are removing one more part to be faulty, or become loose. Therefore, the device *should* have a longer MTTF.
Cheaper Power - A desktop running at 15W instead of 250W? Can you imagine the power savings if you did all of your desktops this way?

If you want to know more about the HP t410, Chris has done a really nice writeup at http://kontrolissues.net

I purchased my first HP Switch!

2012-10-09T22:47:49Z

On Sept 12, I posted a tweet saying that I purchased my first HP Switch. Obviously, this caused some discussion between the HP lovers, and the Cisco loyalists (okay, that's an exaggeration). In this post, I hope to communicate to you *why* I chose to purchase 2 HP 5406's, even though I'm working on my CCIE and I have Juniper EX's running around.

First, let's talk about the current environment. I have an office that is running a Cisco 4510E with redundant 220 power supplies and dual sup's. This switch is acting as both an aggregate switch as well as a collapsed backbone. She's a good switch. No problems (other than both supervisors losing their configs 4 years ago) whatsoever {{knocks on wood}}. It just runs. So, why replace?

1). Maintenance - Maintenance is killing me! I work for a private company owned by an equity group. OpEx (SG&A) is taboo, and if the powers that be could run the business without it, I'm sure they would. Anyway, this single switch was costing me approx $10K in annual maintenance (yes, smartnet).

2). Future proof tech - I think we can all agree, that the 4510 won't be moving us into the future. I can just see it now... 6 months from now, a challenge presents itself, and I answer "We can do MPLS. All I have to do is........ ((sigh)) nevermind."

So, now that we identified why I wanted to get rid of it, let's get back to the question, why HP?

1). Cheaper. - Now, if you know me, I'm all about spending the money if I feel it's worth it. If there is value in it, I don't mind spending more. However, to me, I didn't see the value in the Cisco name anymore. The HP switch does everything I need it to, and it is reliable. How do I know? Because I've been running 2 5406's in Sweden for 3 years without a failure.

2). Maintenance - Anyone want to guess how much HP's maintenance is? Anyone? $0. Warranty covers it all.

3). Simple - No licensing images to choose from. One image has all the features. I don't have to worry about if I'm purchasing the L3-Lite or Advanced Security image. What the datasheet says the 5400 can do, that's what I'm getting. (well, at least, it looks that way).

4). Management and Familiarity - I'm already running HP iMC, which can manage the 5400's really really well. And, as I mentioned, I'm running 2 in another country, so I would like to get more familiar with them.

The only caveat I can put out there? Well, I have actually received them and installed them yet. So, maybe I will change my mind. I doubt it, but it's possible. I just wanted to tell you the reasons behind the choices. My only regret, was that I could not afford a bigger switch running Comware. Do I need it? Not for this case, but I would just really like to dig a little deeper under those covers.

Microsoft Event to SNMP Trap

2012-09-24T17:13:04Z

If you have been reading my posts on SNMP and management, you will see a pattern....... "I hate Microsoft's position on open management". It's mostly because Microsoft will not make a move to SNMPv3. However, there is one thing that Microsoft did really well..... and that's the ability to turn ANY windows event into a trap.

Brilliant!

So, now, no matter who the application vendor is, if an event is created in the Event Viewer, you can translate it to a trap. Think of all the wonderful alarms you can get! It was kinda surprising to me how little this nugget of info was lost. So, I decided to write up an entry about it.

We are always going back to our resolved problems and asking ourselves, "How can we prevent this again", or "How can we see this problem faster?" Or, at least, I would hope you ask yourselves those questions.

Well, more often than not, if it's a Windows application, there is probably a Windows event that is generated prior to the problem happening. Some "symptom events", if you like. Identify those events, and change them into traps.

Once you have a trap going to your NMS, you can alarm and escalate it as anything else.

EUREKA!

Global Department Meetings

2012-09-24T17:14:28Z

This past week, my department had a global meeting, where many department managers and function leaders get together to talk, brainstorm, and learn new things within the company. We should have this every year, but our last one as a few years ago.

The meetings themselves were fantastic and there were some great ideas and interest from the event. But, when you are trying to do your day job, it's a little difficult to find time to build your presentations.I had 4 topics to present:

Service Management and ITIL. Details on our ServiceDesk functions, procedures, and workflow. And the "Why".
Document Management Systems - How they can be used, how we use it, and forward thinking.
VDI / BYOD - What it is, the benefits, and how we will use it.
Oracle UPK - User Productivity Kit - How to use it, and getting it adopted through global channels.

Number 1, 2, and 4 were easy enough for me. Those are topics that 'I get". I've been doing it long enough where I really don't even need a presentation. Put me on stage, give me a mic, and I'll just start talking about them.

Number 3 was hard. It's a new concept for the company, and I don't know enough about it. Sure, I know Citrix back in the Metaframe days. I also know virtualization. However, I do not know Citrix's implementation well enough to answer questions or articulate certain ideas. So, let's start doing research, opening a powerpoint, and brainstorm. Oh, did I forget to mention, that this is during an already busy week, and my presentation is in 12 hours?

Luckily for me, I have access to a new technology called 'Internet" and "Google", which I will call "Intergle". With Intergle, I was able to get the topics, bullet points, and images I needed for the powerpoint. (( Enter excited narrator's voice from a commercial )) And that's not all! With this product, you can even find new job postings, when you get fired from your last one by doing presentations too late!

Overall, it was a great week, and I was sad to see everyone go. Unfortunately, I could not see all of my team, because... well.... someone had to stay behind and support those sites. Regardless, I pray all has safe flights, and happy families to return to.

Thanks everyone, for enriching my life just a little bit more this week!

Cheers / Salute / Salud / Skål

HP eAPI Docs for iMC

2012-11-08T18:44:18Z

I'm always searching for the documentation on HP's iMC eAPI. So, here it is:

~~HP Intelligent Management Center Extended API Guide~~

HP Intelligent Management Center Extended API Guide (new location)

Some of my favorite Mountain Lion Features

2012-08-13T02:39:09Z

Apple has released another Macintosh OSX version, called Mountain Lion (10.8.x). I'm not the only one getting it. Apple released that by July 31, over 3 million downloads were completed, making it "the most successful release in Apple's history". I'll admit, I'm not seeing too many features in this one. Sure, there are lots, but nothing that really affects me per se. There are a small few, which I'll outline here.

They have certainly deviated a little bit from the version naming nomenclature when releasing new versions. Most major releases that you pay for are usually on the left side of the decimal point (i.e. 10.x, 9.x, 11.x). Then, you have the minor releases on the right side (i.e 10.8, 10.9, 10.7). And, many times, you will see the fixpacks or service updates, or maybe even build numbers after that. So, you may have fixpack 1, to update 2 of version 10 (i.e. 10.2.1). Not Apple. Because OSX is version 10, they just keep changing the minor version, even thought they are major upgrades.

Will there be a 11.1? Will that be OSXI? Could it be OSWORLDDOMINATION? Time will tell.

Back to the features (tangent, anyone?):

1). Notifications - Okay. You can't talk about Mountain Lion without talking about notifications. It's just awesome! Before notifications, we had Growl. And Growl is still nice. But, I like that you can pull up notifications at anytime to see past items. With Growl, you really don't have that (or I haven't seen it).

Up until now, most apps were supporting Growl notifications. But, if you look at the Notifications Preferences, only the built-in apps are configured for notifications. So, what are we to do? Use Hiss. Hiss allows you to move your Growl notifications into the Apple notifications. Eureka! At the time of this writing, it's currently in Beta, but try it out! (Thanks to @Th1nkdifferent for this nugget of info).

2). Safari Enhancements - I was a Safari fan. Then, I was converted to Firefox. Firefox took up too much memory, and I went to Chrome. Chrome was faster but still wasn't always accepted amongst the narrow-minded web developer community (hint: don't block us because we aren't using Browser X), but with the new Safari, I've come full circle. (Yes, Internet Explorer was not even mentioned for a reason. DRAT, I just mentioned it!!) There are a couple of enhancements that has made me reconsider Safari as my primary browser.

The first is the search/address bar. Taken from Chrome, you can now both search and type in a full address in the address bar. Do you have any idea how easy this will make for our helpdesk technicians working with a non-techie user?? Plus, it just makes it easier for us. I will admit, it didn't always work for me. I haven't identified the pattern yet, on what is acceptable as a host address versus a search.

The other feature is the tab pinching. If you are in Safari, you can pinch gesture on your touchpad, and the tabs will resize to a carousel that you can swipe to the tab you want. Handy!

3). Embedded Twitter - Finally! A twitter client that is embedded into the OS. Now, that does not mean Twitter is fully functional. Far from it. It does, however, make it easier. And, hopefully, with the API's, a much richer set of Twitter apps on the horizon.

So, what does integrated Twitter mean for you? Well, first, it hooks nicely with Notifications. You can even tweet from the notifications bar. It also makes it easy to tweet web pages, as it's hooked into Safari as well. But, that's about it. Limited, yes, but the possibilities are fantastic. The notifications alone are great.

4). Messages (formerly iChat) - So, take iChat, and enhance it with iMessages, screen sharing, and file transfers, and you have iMessages. It's really fantastic. I've already been using it more than I should. With Jabber, Google Talk, and iMessages support, I pretty much have everything I need there. I've already helped out my wife while at work. No Webex, join.me, or GoToMeeting for me. Fire up iMessages, and you're golden!

Plus, because it adds iMessages support, I can chat with my wife's Jabber account transferring files and chatting. Then, she goes offline, and I just switch to her iMessages and start texting her on her iPhone/iPad. I didn't even switch windows.

Can't open Parallels after Mountain Lion update

2012-08-05T14:31:27Z

I just upgraded my Mac to Mountain Lion. So far, I like it. Especially the integration of Twitter into the OS, and the enhancements to iChat (now called Messages). However, there was one thing that bothered me, when I tried to open Parallels.

First off, my Parallels dock icon showed it cannot be opened:

You may also get the error message "You can't use this version of the application Parallels Desktop.app with this version of OS X."

According to Parallels website, you need to re-install using the latest build of the app.

Installing HP iMC with remote SQL Server

2012-07-10T04:00:30Z

When installing HP Intelligent Management Center, you have a choice on where to store your data. If you are on a Windows system, your options are SQL Server or Oracle. When installing on a Linux system, your options are MySQL or Oracle.

Initially, my install was on a Linux system (CentOS), for reasons beyond the scope of this post. I installed MySQL and setup all the configurations as defined in HP’s “MySQL 5.5 Installation and Configuration Guide (for Linux)” guide. Everything worked great. MySQL server was up and running, and I was able to connect to it with other tools. But, iMC would not pass the database connection check on the installer. I could not get past it, and I was forced to use Windows. By the way, even though CentOS is technically RHEL, just without the support, iMC Tech Support will not support it.

So, Windows (cough) it is. It’s not that Windows is terribly bad…. I just prefer Linux/Unix when it comes to Network Management Systems (and for Java-based software). Hmmm… that’s my 2nd note about Linux…. I’m thinking I need to write something up. :)

Since I don’t use Oracle, I can only use SQL Server. Now, you have 2 kinds of SQL Server. You have SQL Server Express, which is a freebie version giving you 10GB of storage per database. However, not only are you limited to the database size, but installing locally means your Disk IO may not be utilized best. But, it’s easy, and works for small to medium size management needs. I chose to use our current SQL Server which doesn’t have storage limitations and is tweaked for best usage (plus, I will be using heavy data collections and reporting).

Caveat #1 - I didn’t see it in the docs, but thanks to Twitter, @netmanchris tips me to installing SQL Server Tools. If you don’t, then iMC will assume a local install of SQL Server Express, and it won’t ask you about a remote SQL Server.

Caveat #2 - Make sure the SQL Server you are using is setup for Mixed Mode authentication. I could not get the installer to use a Windows login for setup.

To make sure you are prompted for a remote SQL install, choose a “Custom” installation when installing iMC. It seemed that choosing “Typical” assumed a local install. I could be wrong about this, as I did not have SQL Tools installed, but the docs say to use “Custom”, so I did.

On one of the screens, you will be prompted for the data location shown below. The installer requires SA permissions because it creates logins and databases.

This is the directory on the remote SQL Server that will host the databases. You will need to create the directory.

Caveat #3 - If you are using a remote SQL Server, chances are, you are storing your logs and db’s on seperate disk arrays to keep Disk IO down. Since you are not creating the DB’s yourself (the installer is), you are bypassing the standard SQL policy on file storage. This means that the db’s and logs are put in the same directory. This is not efficient, and you’ll need to move them.

To move your log files, you’ll need to stop iMC services, detach the databases that were created, and move the log files (ldf) to the other drive. When you attach the database, it will say “Not Found” next to the log file (Duh, you just moved it). Change the path to where you moved the log file, and say OK.

You will have to do this on each DB that was created. If you are installing just about everything, the DB names are:

aclm_db
config_db
etl
icc_db
invent_db
monitor_db
perf_db
report_db
reportplat_db
syslog_db
unba_master
unba_slave
vlanm_db
vnm_db

And there you have it. Remote SQL Server for iMC. Once you understand it, it kinda makes sense. However, if I were to have a feature request list:

1). Allow a prefix to be added to the DB's. If this is a remote server that is using a shared instance name, then it's best to identify them from all the other DB's. Prompt for a "Prefix" to be added with an underscore. Such as: Prefix: "IMC". Thefore, the DB names will have IMC_name_db on them.

2). Don't assume a local install unless SQL Server Tools are installed. Ask if it's a local or remote. If it's remote, then prompt a dialog box saying SQL Tools could not be found and are required. Of course, SQL Tools can also be installed using SQL Express, so, maybe just install them?

3). Allow for the user to define BOTH the data and log file locations.

HP Intelligent Management Center - Service Health Management

2012-06-11T18:42:31Z

20 or 30 of us walk into the next HP Discover 2012 break-out session titled “Understanding Service Health Management in iMC”. Les Stuart begins the presentation with the overview of SHM. Chris Young, Solutions Architect for HP walks to the front and asks a single question, “How healthy is your network?”.

Now, as you can imagine, most didn’t know how to answer it, and after waiting about 15 seconds of mumbling and crazed looks, people began giving different answers. That’s to be expected, since everyone determines that differently, depending on your responsibilities. The Server guys think the network is healthy by a simple up/down status on their servers. Voice guys base it on QoS and Jitter. Network guys, something different.

With a smile on Chris’ face, knowing those were the answers he what he was expecting, he then proceeded to show how flexible the SHM module of HP’s Intelligent Management Center can answer exactly that question, no matter who you are. Whether you care about SLA’s, thresholds, or simple metrics, SHM is the iMC standard for reporting on the metrics you care about rolled-up into a single view, number, or graph.

The reason behind SHM is simple: iMC is already gathering statistics on your network. Whether it is jitter, QoS boundries, up/down status, or bandwidth, iMC already has that data. Why not tap into it to give you better reporting? If iMC can get it, you can use it to build your “health”. And let me tell you…. iMC CAN get it, whether it is via SNMP, WBEM, sFlow, syslog, SSH/Telnet, or Python/Jython scripts.

It all boils down to KPI and KQI. These are standard Business Intelligence/Scorecard/Dashboard terms meaning Key Performance Indicator and Key Quality Indicator, respectively. You build your Key Performance Indicators from the desired metrics (for example bits in/out), and roll them into a Key Quality Indicator. You then use your KQI’s to build an SLA/Health Management Report.

“But I can get that with MRTG”, you say. Well, yes you can. But it does not have to end there. You can roll up a number of KPI’s (say, bits in/out, up/down, and TCP errors) into a KQI. You can also roll up server latency and response times in a different KQI. Then, using both KQI’s build an SLA Health Report. The power of SHM is not just simple reporting, but rather, using the current data-collection facilities that iMC already has, and build a dashboard from it. See image below (click for larger):

When I first started writing about SHM, the features and details were far beyond what I single post can offer, so this will be broken up into multiple-posts. Stay tuned!

Interview of HP's Saar Gillai on SDN

2012-06-09T18:02:07Z

During HP Discover 2012, I've been asking various HP divisions on the plans HP has for an OpenFlow controller. The reason? For the last 18 months, I've noticed that every HP event mentions SDN and OpenFlow. Why? Just adding support for a technology is not a big deal. One, maybe two announcements, and you are done. Why so much talk about it? Surely this means that HP has big plans for it.

The answer I get back is, "What would you do with one today?". Well, nothing, TODAY. The enterprise doesn't really have a need for it yet. But that's just redirecting my question. I don't care what the controller looks like, whether it's in iMC, an appliance, or a hybrid. So much talk around SDN.... surely HP will do more, right? Otherwise, why keep talking about it?

So, where do you go now, HP? The controller would be the next logical step. Yes, HP has done some very cool things called "VAN"s (Virtual Application Networks) within the iMC, which I will post about soon. But, that's a small nugget of the overall SDN pie.

During HP Discover 2012, Stu Miniman (@stu) from wikibon.org interviewed Saar Gillai on the concepts and direction of SDN as it is seen from an HP perspective. While he does not actually answer when, he does say "Stay Tuned". Check it out below: