Keeping Your Warp System Healthy and Happy

A Bit of Prevention Saves Days of Pain

You can choose to be good at troubleshooting or you can avoid the need for troubleshooting.  A little prevention goes a long way.   For some reason, if you have backups and recovery procedures in place, you rarely need to use them.

Hardware Maintenance

Todays' hardware is very reliable.  If it's alive when you install it and does not die young, it will probably outlive its useful life.  All the old, slow systems still running are evidence of this.  Nevertheless, there are a few things you can do to prevent hardware problems.
 
• Heat.  Get up close and personal with your computer case.  Know what it's normal temperature is.
• Fans.  Know the sound of your fans.  Know how many you have.  If the sound changes, it might be time for a trip to the computer store.
• Dust.  Dust will make your system run hot and, worst case, can cause shorts and intermittent electrical failures.  Check the fan inlet every once in a while.  If it's dusty, open the case and blow out the dust.
• Power.  In the days of rolling blackouts, it's a good idea to know the quality of your power.  If you are subject to frequent power outages, a UPS is a wise investment.

Backups

• If you don't have them, you will need them sooner rather than later.
• Chose any backup medium you are comfortable with.
• Test your restore procedures.  Do not assume they will work when you need them.
• Backup your Master Boot Record and Partition Tables.  This may save you a full restore someday.  The  Graham Utilities have tools for this as do the  GammaTech Utilities.  Unfortunately, these tools do not yet support backing up the LVM signatures.

Desktop Maintenance

A properly cared for Desktop can be very robust and resilient.  However, the WPS has tendency to not clean up all references to deleted objects.  If you do no use third party tools to remove this obsolete data, over time the debris will pile up and cause problems.   The symptoms are the Desktop will appear to get slower over time and will tend to crash more often.  These problems are easy to avoid.  Just use a WPS cleaner on a regular basis.  There are several available:
 
•  Henk Kelder's Checkini (WPTools)
•  Carsten Arnold's Cleanini
•  Unimaint


Each of these tools does the same basic job.  However, each has cleanup features the others do not.

The Desktop should be backed up separately from the rest of the system.  Both WPTools and Unimaint include WPS backup and restore tools.  There are other options.   As always, be sure to test your restore procedures before you need them for an emergency restore.  You should also maintain at least one OS/2 Archive backup.  The Desktop restore utility that comes with Warp is are not as reliable as the others, but the installer creates a Desktop backup image when the system is installed, so you might as well maintain it.  Rather than using the settings folder to create new images, use archinst.  This is what the installer uses to create the initial backup.

Config.Sys

• Keep some sort of config.sys history.
• Keep os2\boot\config.x in sync with config.sys as needed.
• Retest Alt-F1 boot to command line after any driver install.  Update config.x as needed.
• Enable application trap logging with:
SUPPRESSPOPUPS=X
replacing X with the drive where you want POPUPLOG.OS2 written.
• Suppress application restarts with:
SET RESTARTOBJECTS=STARTUPFOLDERSONLY
• Memorize the CTRL-SHIFT-F1 key combination.

Applications

• Know what you have installed.  Warp even includes tools to create an installed application database (Netfinity).
• Have a basic understanding of where each application stores its custom settings.
• If the application is not stable, consider a specific backup procedure the application settings.

OK, it's broken. What now?

• No Response to Power On
• Check power.
• ATX cases are different.  Sometimes the power switch is ineffective.  You need to unplug from the wall.
• Sometimes, pressing reset quickly several times works.
• POST Fails
• Check to code.
• Reseat all the cards
• Reset all the socket chip.
• Plug all cards are retry inserting one card at a time.
• BIOS Can't Find Disk Drive
• Check cables.
• Make sure disk spins up.
• Check power to drive.
• Warp Says It Can Not Operate Drive
• Use the BIOS Verify Drive feature to ensure disk has no new bad sectors.
• Figure out how you changed your partition tables.
• Lost Desktop or WPS Will Not Start
• Restore Desktop from backup.

There are tools such as Object Recreator that will restore some individual lost objects.  However, a full Desktop restore is usually takes the same amount of time and cleans up the problems you don't see.
• Application Will Not Start
• Try to remember what you just changed and change it back.
• If you installed a new application recently, maybe there's a LIBPATH conflict.  You can use chk4dlls,   (chkdll.zip),  chkdll32  from IBM, or pmdll v2.3. to check that all the required DLL's are available and that they are getting loaded from the application DLL directory.

Pmdll has a nice PM interface, but has problems with long LIBPATH lines.  A new version will be available shortly which corrects.
• Sometimes application specific setup files get corrupted.  These can usually be restored from a recent backup.
• Peer Can Not See Other Peers.
• Make sure Domain Names or User Names have not changed.
• If the other Peer is a WinXX box, make sure that LMAnnounce has not been turned off.
• TCP/IP Connection Problems.
• Check lantran.log to ensure that the drivers are loading correctly.
• Monitor the Switch and NIC LED's to ensure data is flowing both directions.
• Monitor LAN interface activity with iptrace/ipformat.  This works for both direct and dialup connections.
• If you have dialup problems, use DOIP's debug option to trace line activity.

Built-in Diagnostic Tools

Warp contains a variety of built in diagnostic tools.  Some, we have mentioned already.   The ones I use most often are described, briefly below.  Some have sufficient functionality to warrant a presentation in themselves.  If you hunt though the folders on your Desktop, you will discover some I have not mentioned.
 
• POPUPLOG.OS2.
This is where OS/2 stores the Process Dump summary.
• Process Dumps.
Warp has had the ability to dump failing processes since before the GA version.   With recent versions of Warp, this facility has been significantly enhanced.  Now, it's possible to tailor the dumps and to dump on demand.  The documentation is at \os2\system\ras\procdump.doc.
• System Dumps.
Warp has long had the ability to dump the kernel image.   As with Process Dumps, this facility has been significantly enhanced over time.  Now, it's possible to tailor the dumps and to dump on demand.  The documentation is at \os2\system\ras\procdump.doc.  There is also a  Trap Dump Tutorial  published in the January, 2001 SCOUG Newsletter Mr. KIA column.
• Iptrace/Ipformat
Iptrace captures activity on your LAN interface.  Ipformat convert the binary dump file to a readable image.  This trace is useful for determining exactly what's failing when you have Internet connection problems.
• System Trace
The System Trace allows you to capture detailed information on almost every system action.  Much of the information requires known of kernel and subsystem internals to interpret.  However, there are some items that can provide useful information.  One example would be to discover the file name when an I/O operations fails and they application does not report the file name.
• System Error Log
• The System Error log reports exception information passed to the kernel from applications.  The best use of this for end users is to monitor the health of the WPS.

Other Diagnostic Tools

There are many other handy tools available.  I've even written a couple myself.  One is:
 
•  Trap Screen Extraction Tool   This is a REXX script that I wrote to overcome the need to write down the Trap Screen registers.  It extracts the data from a dump image.   It does not require a full dump.  You can dump to diskette and kill the dump after the first diskette.

Post Mortem

This just touches the surface of what's available.  If you have any specific questions, you ask them on any of the SCOUG mailing lists.

SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group. OS/2, Workplace Shell, and IBM are registered trademarks of International Business Machines Corporation. All other trademarks remain the property of their respective owners.