Sonicwall GMS 5.1

Sonicwall has a management product called GMS which is at version 5.1 as of today. This web-based utility attempts to keep track of your Sonicwall products via syslog and provide a central point for all administration. Unfortunately, the default install relies on MySQL and the default parameters are not effective in keeping the service stable.

One of my biggest gripes with the product is how it does heartbeat counts through Syslog. Syslog uses port 514 and is UDP traffic which is not the most reliable if something was to go wrong out on the WAN. The syslog traffic is then fed into the database platform and unfortunately MySQL does a really bad job of staying up long enough to capture everything.

If you look at the documentation for 5.1 there is a section detailing how to go about installing SQL Server 2005 and above which is apparently supported. However, unless you choose Multi Server install during the initial setup you will not be prompted for a database location other than using MySQL. This is obviously a big misstep on the part of Sonicwall to leave out something like this when most shops of any decent size will want to have a central database server for all of their apps. The documentation is very dry and does not give any detail around the SQL Server 2005 piece which is disappointing. Another problem is that there is no way to migrate the database using built in tools from MySQL to SQL Server which means you have to re-enter your data once you make this migration.

Those are my major gripes about the product but once you get it set up and functioning in an enterprise fashion it does seem to work well. Configuration changes are pushed out to your devices as you would expect so that you don't have to log in to all of your devices.

Specialists vs Jack of all Trades

I am a big proponent of having a lot of knowledge about various technologies rather than be specialized in one particular area. When I set up an IT staff, I want people that can do multiple things and be flexible in terms of coverage. This allows you to hire less people but get more done. To me, this has been pretty much common sense. Unfortunately, larger organizations have focused on hiring many specialists that do one thing very well which is horribly inefficient. At some of my previous jobs, there were IT staffs of over 1000 people that seemed to never get anything done. The best IT staffs I worked on were of less than 10 people where a weak link was singled out very quickly and dealt with. In these types of staffs you can see where you are weak and make adjustments very quickly. In large organizations it is easy to hide amongst the masses when you basically do nothing to contribute to the overall IT team.

A good article about this topic came up at Interop and I want to share it here with you:

http://www.networkworld.com/news/2009/052009-interop-job-skills.html?fsrc=netflash-rss

Solarwinds Orion

Solarwinds had an IPO today which is significant for a couple of reasons. The economy as we know is not in good shape so its important to see businesses have the ability to go out and raise capital. Any IPO, especially a technology related company is a great sign for our industry. On a more personal note is that I have been a user of Solarwinds products for several years now at various jobs throughout the years and really enjoy their ease of use and low cost of ownership. One of the best tools that Solarwinds puts out is the Orion monitoring tool.

Orion uses SNMP to monitor your devices on your networks, whether it is a switch, router, firewall, server, etc. Alerts can be set up to send emails when certain thresholds are met, such as volume size or memory utilization. This type of monitoring is much more proactive than just telling me if a server or device is up or down and gives a good overall view of the health of your infrastructure. The GUI is web-based and has a great reporting engine which is pulling from a database so you can see historically the uptime of a particular server, T1, or any other device you are monitoring over time. This is great to show disk usage and how it is growing over time in order to justify additional disk space.

The software is licensed on a per node basis. This means that if you want to monitor all of your interfaces on every switch plus all of your servers then your node count will be very high. Of course, you don't need to monitor every interface on your network. Concentrate on your high value interfaces such as uplink ports on your switches, your router interfaces, your firewall interfaces and critical servers to keep your node count down. It will take some planning but if you lay out what is most critical for you then you can dramatically cut the costs of this monitoring tool.

CICS and SFS

One of the most annoying issues with CICS on a Windows/Unix platform is SFS. CICS uses SFS to access its "database" which is nothing more than a bunch of flat ascii files. This is really not a database but can cause you a lot of grief if your CICS region is shut down improperly and SFS has a lock on a particular file. If your SFS file system gets hosed, one of the easiest ways of dealing with the problem is to simply blow it away. Actually, I recommend this to clean up your SFS database if you are experiencing any SFS related errors in your console.log. Remember that once you blow away your SFS database and re-create it, you will have to recreate your region as well. Once you re-create a region, it of course will be blank until you reload your definitions into the database. There will be no listener or any program defitions, among others. Most people make the mistake of creating a region with the cicscp command but not loading any information in it. By default a CICS region is very dumb and has little useful information until it is loaded with anything specific. In a later posting I will go over CICS tuning and how you can maximize the performance of your CICS regions with a few simple tweaks.

Converting from Dynamic to Basic disks

One of the worst blue screens of death you can get is a stop error on boot 0x0000007b which is an inaccessible boot device. One of the reasons that can cause this is trying to convert your boot disk to a dynamic disk which can cause you these types of problems. Unfortunately, there is no supported method of switching a disk from dynamic to basic and you need to hack the disk into thinking it is basic. This situation came up this past weekend when attempting to use Acronis Disk Director to increase the size of a particular partition on the system disk. When the server rebooted, the stop error happened. When attempting to boot into safe mode or last known good configuration, the same stop error occurred. Fortunately, we found an article at:

http://thelazyadmin.com/blogs/thelazyadmin/archive/2007/01/17/Converting-Dynamic-Disks-Back-to-Basic-Disks.aspx

that talked about how to modify particular sectors on a disk (01CO for the C partititon, 01D0 for the D partition, etc) and change the third byte from a 42 (indicating a dynamic disk) to a 07 (indicating a basic disk). The tool described in the article only works inside of Windows and does not work from a boot disk. Fortunately, a copy of Hiren's boot cd includes a copy of Norton Disk Editor which allows you to boot your dead system, modify the sectors for each partition, write the changes back to the disk, and then boot. This worked flawlessly and allowed the system to boot properly. Just make sure that you modify the byte for each partition on the disk that you want to convert back to a basic disk and your system should be fine.

Training

In these tight economic times, one of the first things to be cut from any IT budget is training. Unfortunately, I view that as a big mistake because when times are bad, you want to be able to get the most out of your employees instead of hiring a FTE or even bringing in an expensive contractor to fill a gap. Paying for your employees to travel off site and attend a training class may not make the most sense because of your staff size and the work that they usually perform during normal business hours. Training should be customized to fit the needs of your staff and situation so that it makes the most sense to all involved.

As a former Senior Technical Consultant for Mincom, I have conducted several training classes both internal and external to Mincom and can give the training that you require based on the needs of your environment. Not only can we deal with Mincom software issues, but we can also talk about best practices for your entire infrastructure with Mincom products. Training can be customized to fit your needs and not a one size fits all approach which will not give you the best use of your investment. Please contact us today to get an idea of how we can tailor a training experience that will fit your business.

Mincom Ellipse

I used to work for a niche ERP vendor some of you may have heard of named Mincom. My role with the company was a Senior Technical Consultant who specialized in installations, performance tuning and infrastructure design issues. As of the 12th, I was laid off due to the current economic conditions and the poor performance of the company in general. I was the last one hired on the team and the first to be let go. This can only be a benefit because I have a lot of knowledge and practical experience in their software which I can now use to help other companies who may be having issues or problems. It is quite expensive to upgrade your ERP system, not including the time and effort required for training and infrastructure upgrades so you have to ask yourself what exactly are you getting for your investment? Maybe your CICS performance is poor and not properly turned. Is your Ellipse infrastructure for 5.2.3.7/8 set up properly? These are issues that are not covered by your support contract and unfortunately can be very costly to have Mincom diagnose these issues. Now that I am no longer with the company, I can offer these services remotely to future clients. Please let me know how I can help you get the most out of your Ellipse environment.

Backup exec problems

One of the most popular tools for backing up your Windows server data is Backup Exec. Unfortunately, it can also be one of the most frustrating pieces of software to use due to its convoluted error messages.

A couple of scenarios that have came up recently involve having the ability to move the Media Server database from the default install to a new server or even a full blown SQL Server installation. The utility appropriately named "BEUTILITY" will accomplish this task for you which turns out to be relatively painless:

http://seer.entsupport.symantec.com/docs/275658.htm

The second issue is somewhat of a pain in that it deals with SQL Server backups. SQL Server has a built in default path that tells it where to backup databases to when you kick off a backup. This can be found in the registry at:

HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Microsoft SQL Server\MSSQL.1\MSSQLServer\BackupDirectory

Change this to a location on your file system where you can grant the appropriate NTFS permissions for your Backup Exec account to access. For example, change this to "d:\sqlbackup" and then restart your SQL Server services. Once that is done, your backups should be fixed.

For more information, see the following link:

http://social.msdn.microsoft.com/Forums/en-US/sqldatabaseengine/thread/021c7aa5-4a8a-4bbb-8ff0-fe6b03920aae

The economy

Unfortunately, no business or profession is immune from the economic turmoil that is going on in the world right now. No one seems to be investing or hiring for the future. IT is usually looked at as a quick place to make cuts that really don't contribute to the bottom line directly, but these decisions are short-sighted. The productivity gains of a good, solid network that is not prone to viruses, outages, or other issues gives the workers who contribute to the bottom line the ability to do more with less. Investing in your IT infrastructure may seem pointless in bad times economically but imagine the cost if your critical applications are down and no one can access the financial systems.

Of course, throwing money blindly at problems is not the solution. Do your homework and show the metrics and return on investment that spending the money will bring your company. Gather the data and show performance statistics that are easy to understand and explain. If you don't have the expertise, consider outsourcing or contracting someone with the skills to come in and take a look at your environment that will provide an independent and non-biased assessment of the current issues. Having this outside analysis will only help you convince your superiors of your findings and show the need for further investment.

PMT and Websphere 6.1

I did my first install of Websphere 6.1 the other day on HP-UX and ran into a very strange problem. Profiles are now managed through the PMT, the profile management tool. After doing the initial install, I started the PMT through the first steps wizard and the PMT launched, but the screen was blank. I didn't see any buttons or anything to click on, which was odd since the websphere install itself worked fine with my X server (I prefer the free XMING, or you can use Cygwin inside of Windows). Due to security restrictions on the VPN I was connected to, I was unable to try the built in X server on my Macbook Pro, so I was stuck inside of Windows.

I did some googling and found this article on an IBM forum:

http://www.ibm.com/developerworks/forums/thread.jspa?messageID=14238812

Apparently this is a known issue and the only X server on Windows that works as of today is Exceed, which of course isn't free. I had to download the trial, install it, and then magically I was able to see the buttons to do a GUI based profile creation. Not cool, and I hope IBM fixes this fairly soon to work with other X servers.