CallManager Trace: Update

When downloading trace files for callmanager, you may notice that you have two sets of files: SDI and SDL. I set out to get the definition of these acrynyms, and found the following:

SDI: System Diagnostic Interface
SDL: Signaling Distribution Layer

I think this helps to explain why you want to look in SDL for tracing call info!

Google Search Appliance Pricing

While doing research on the new version of the google search appliance, which is the beefier version of the google mini, I got my hands on the price sheet for the various models and licensing price points. While it does seem pretty steep, when you have a corner on the search market you can charge what you want!

----------

The following models include a two year license fee which covers hardware,
software, technical support, and product updates.*
Pricing is based upon the total number of documents (or unique URLs) in the index.
Google Search Appliance Models
Hot Backup and Premium Support
Google Search Appliance Price Sheet
Model
Documents Indexed
Price
GB-1001-500K
Up to 500,000 documents
$30,000
GB-1001-1M
Up to 1 million documents
$50,000
GB-1001-2M
Up to 2 million documents
$100,000
GB-1001-3M
Up to 3 million documents
$150,000
GB-7007-5M
Up to 5 million documents,
includes redundant appliance
$250,000
GB-7007-10M
Up to 10 million documents,
includes redundant appliance
$400,000
GB-8008-15M
Up to 15 million documents
$550,000
GB-8008-30M
Up to 30 million documents
$1,000,000
Model
Price
Hot backup / development unit for any GB-1001 model
$20,000
Hot backup / development unit for any GB-7007 model
$50,000
Hot backup / development unit for any GB-8008 model
$250,000
Standard Support for all models
Included
Premium support for GB-1001 (includes a hot backup)
$25,000
Premium support for GB-7007 (includes a hot backup)
$125,000
Premium support for GB-8008 (includes a hot backup)
$325,000
Choose Standard Support if access to support during business hours only is sufficient.
Choose Premium Support if 24X7 access to support for Critical (Severity 0) issues is required. The Premium Support option ensures maximum uptime for mission critical applications, by including the following features:
• An identical failover appliance is provided. This secondary appliance is required to be deployed in tandem with the primary appliance, crawling identical content.
• 24X7 pager coverage for Severity 0 issues
Google recommends obtaining the Premium Support, along with additional development units (as needed). This architecture ensures a separate development environment and maximum support on the production environment.
Support levels defined at http://support.google.com/enterprise/terms
*For customers who wish to upgrade to a higher license volume tier during the two year term, Google simply charges a prorated monthly fee for the remainder of the term.

Understanding Cisco CallManager Traces (CUCM)

Whether you call it Call Manager, Unified Communications Manager or CUCM, everyone agrees that knowing what is going on under the hood when there is a problem is a must. Having had the pleasure of calling Cisco TAC numerous times to find out what happened to a call I finally have the process for gathering traces down. It still takes quite a bit of experience to tell what is going on sometimes and that’s why I’ll still be calling TAC to help decipher the traces. Simply getting the traces saves a lot of time, so I’ll outline the process. The last issue I worked on had to do with debugging ISDN/Q.SIG messages, so I’ll refer to those throughout.

Step 1: Turn on Traces in CallManager
Open up your CallManager Admin page and navigate to Serviceability. Then go to Traceà Configuration, select your publisher and choose the CallManager Service. Make sure Trace is on, and that debug trace level is set to Detailed. You will also see a lot of options with checkboxes. I normally leave these to the defaults, but if you are looking for something specific, go ahead and set these. The ones I am normally sure to check are the PRI trace, MGCP trace and All Phone device traces. Now make sure that the “apply to all nodes” is also checked and click save.

Step 2: Download and Install RTMT (Real Time Monitoring Tool)
This tool is accessed from the normal CallManager Administration page. It can be found under ApplicationàPlugins. Download and install this tool. Make sure you can run the RTMT tool and log in using the admin credentials you use to log into the web admin page. You may see a “select configuration” dialog box. You can cancel this. It is for setting up a default set of windows of CallManager performance data. We just want the traces. Choose ToolsàTraceàTrace & Log Central.

Step 3. Place Your Test Call
Now that you have RTMT loaded up, you need to place your test call or create the condition you are trying to capture. Go ahead and do this, twice even so that you know you can find it. Wait ten seconds or so once you are done with the test so that CallManager has a chance to write the trace files.

Step 4. Collect Trace Files
Now go to your RTMT window that you pulled up in step 2, and double click “Collect Files”. Find CallManager in the list, and click the checkbox for “All Servers” since you may have a situation where a phone is registered to one server and the gateway is registered to another. This is good info to have beforehand so your search for the trace is a bit easier. Click next. The next option is to download system logs. These are useful for system issues, but not necessary for trace file collection. Skip this by clicking next. Now for the file options, choose relative range and go back 5 or 10 minutes. Make sure whatever you choose is sufficient to capture the test call or error condition. Trace files contain a surprising amount of data, of which more is not always a good thing. Choose to zip the files and choose browse to select the local directory you want to download the files to. Now click finish. Depending on your relative range this may take some time.

Step 5. Download and Install Triple Combo
Ok, so you could dig right in and look through your trace files for that test call. That’s what I used to do – until I found Triple Combo. This app helps to filter and display CallManager traces, ISDN Q931, H225, H245 – even CCAPI, Q931 and MGCP debugs straight from your router. It also can decode ISDN facility messages. I haven’t found any easier way to do this.

Example:
My particular issue was that I had a Q.SIG trunk from my CallManager environment to a PBX. I had calls going from CallManager to the PBX, and when no one answered, the call got “pulled back” and went to voicemail on the CallManager side. It was not apparent was who was responsible for the Call Forward No Answer (CFNA) behavior. While we had a hunch, we needed the traces to prove it. I was able to get the trace file from CallManager, and found that an ISDN Q.931 facility message was received by the gateway from the PBX over the Q.SIG trunk. I was able to right-click the facility message, choose decode, and saw that the PBX was sending the call CFNA back to CallManager and giving the voicemail pilot number. Now I have the ammunition to go to the PBX vendor and prove that they are the cause for this particular issue.

Now that I (hopefully) have you convinced, install Triple Combo: http://www.employees.org/~tiryaki/tc/ Also download the MT.exe executable. This is required for deciphering the facility messages. http://www.employees.org/~tiryaki/tc/mt.exe
Triple Combo doesn’t install. It is a standalone executable, so make sure that it and MT.exe are in the same directory.

Step 6: Drag Trace Files into Triple Combo
Your trace files are zipped, so you will want to unzip them. The reason for this is because you can use the Windows search feature to look for text in the files if they are unzipped. Instead of searching for a filename, you can search for a word or phrase in the file. This can narrow down which trace file you are going to analyze. The files also need to be unzipped for Triple Combo to analyze them. Once you know which files have the test call, drag them into Triple Combo. In my case, I looked for my Q.931 facility messages, then right-clicked to decode the facility message in the lower viewing pane. There are many ways to filter the output, so I will leave that to you. You’re on your own from here! Kudos to Murat Tiryakioglu for creating this software.

Brad Smith

Changing Terminal Server/Citrix User's File Extensions

I recently had an issue where users on a terminal services session (Citrix) could not change their own file extensions. They could "open with" but like typical end users - this is too much work. I set out to find a way to set this in a more reliable fashion. Of course, if you "open with" you often get the "always use the selected program to open this kind of file" but for whatever reason this sometimes doesn't work. Below are instructions for setting the file association on a per user basis, manually. To do this on a per server basis, see the Microsoft instructions on how to accomplish this.

Procedure for changing someone’s file association in Terminal Server:

Most info is here: http://www.brianmadden.com/content/article/Creating-Custom-File-Associations-to-Support-Side-by-Side-Applications
1. Run “ftype > c:\ftype_output.txt”
2. Look in c:\ftype_output.txt for your program, and make note of the associated name, e.g. Word.Document.8
3. Log in as the end-user, and launch regedit.exe
4. Navigate to HKEY_CURRENT_USER\Software\Classes
5. Create a new Key under Classes. Name it whatever your extension is e.g. “.txt”







6. Edit the “(Default)” REG_SZ value under the newly created key and put in the associated name you noted in step 2. In this case it is Word.Document.8









7. You should be able to test the new functionality immediately. No logoff/logon needed. In my example, files with the .txt extension open with Microsoft Word 2003.

VMware Consolidated Backup, Scripted

I was recently tasked with automating the backup of our VirtualCenter and ESX environment. More specifically, we needed to backup our virtual machines (VMs). When you buy ESX and Virtual Center, you automatically get VMware Consolidated Backup, or VCB. There are framework tools you can install on a windows box to script the vcbMounter command, but I luckily found an even better command: vcbSnapAll. This command can be run from any ESX host (we are on version 3.5) and backs up all VMs that are in VirtualCenter. All of our VMs are stored on a single SAN, but we want to backup to a secondary VMFS partition in case the SAN is completely lost. Below is the syntax I used.

vcbSnapAll -u backupuser -p mypass -a any: -r /vmfs/volumes/MY_SECOND_STORE/

This backs up all VMs listed in VirtualCenter, powered on or off. You can check the syntax options for vcbSnapAll to see what else you can do. There aren’t too many options. If you want to get really granular, you’ll need to write a custom script and use vcbMounter. When vcbSnapAll runs, if you check the running processes, you’ll see that the command is just a wrapper for vcbMounter. Anyway, I then wrote a little script to make room for a second backup. You can’t run vcbSnapAll a second time unless you move the backup files it created out of the way:

#!/bin/bash
#Brad Smith 05/07/2008
#
#This is to clear the previous *OLD* backups
rm -rf /vmfs/volumes/MY_SECOND_STORE/previous_backup/*
#
#
#This is to move the previous backups out of the way, but keeping them in case these backups don't work
mv /vmfs/volumes/MY_SECOND_STORE/* /vmfs/volumes/MY_SECOND_STORE/previous_backup/
#
#
#This actually backups the VMs to the secondary storage
vcbSnapAll -u backupuser -p mypass -a any: -r /vmfs/volumes/MY_SECOND_STORE/

The last step is adding the script (which I created in /root/scripts) into crontab. For you who haven’t much experience with cron, check out this link. It isn’t too confusing. Anyway, here is the line I added in /etc/crontab:

01 04 * * 6 root /root/scripts/weeklyVMbackup

This kicks off the backup at 4:01am, on Saturday morning, and runs as root. When the script completes, check /var/log/vmware/ for the result of your backup. Additional scripting should be done to send these log files if they fail! Too lazy for now to do this.

As always, post comments or questions.

Stupid IPCC Express

This application is so convoluted and confusing, but it is a necessary evil. This will likely be the first of many posts on how to do basic things. Of course with IPCC Express, nothing will be intuitive.

Today's lesson will be: How to make a normal agent user into a supervisor user. All instructions apply to version 4.0(5) but may be useful in other versions. The agent logs into the "Cisco Agent Desktop" using their callmanager userid and password. Not pin, password. Agents use this application to do their daily work. If they become a supervisor, they may want to feel important and log into the "Cisco Supervisor Desktop" to do supervisory things. First, you must log into the CRA/IPCC express admin page "http://yourservername/appadmin". The username is usually CRAAdmin (may be case sensitive). If you don't know your password, good luck. Ask someone who does. Maybe I'll cover password recovery in another post. Once you log in, navigate to tools->user management. Find your user/agent on the right pane, and move them to the left. Check the box that reads "supervisor" and update. Now the user is a supervisor. They need that all important "Cisco Supervisor Desktop" software now. This can be found by navigating within the CRA/IPCC Express admin page to Tools->Plug-ins and then "Cisco IPCC Express Desktop Product Suite". Download and install the "Cisco IPCC Express Supervisor Desktop" software. Once this is done, you'll want to launch the software and log in. Here is the gotcha: you log in as the same username as you would as the agent, but the password is NOT the same. The password to log in as a supervisor defaults to the same as your login id. In other words, if your agent/username is 1234, then to log in to the "Cisco Supervisor Desktop" software your password will also be 1234. Once you log in, you can change this password.

There you have it! Please post any questions or comments.

Safari for Windows

Apple Software Update ran today, and instead of telling me the usual "a new version of iTunes is available" - it said "a new update is available - Safari". What!? I vaguely remember hearing that Safari had a beta on windows, but now it is really here. I'll have to try it out. Look out, Firefox!

---UPDATE---
Safari on Windows sucks so far in my opinion. It is slow, which is my main complaint. Other issues include compatibility with my trackpad scrolling feature, and the fact that new windows pop up in new windows and not tabs. I couldn't find a way to fix these.