Calculating System Uptime
Last Modified: 12 Jan 2007 (Prior Update: 06 May 2004)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*** PLEASE NOTE: Link(s), If Provided, May Be Wrapped ***
Many admins erroneously believe that weekly reboots are
a best practice when dealing with NT/2000 servers. It
is possible to keep an NT4/2000 server running for weeks,
months, or quarters at a time, only taking them down for
hardware maintenance, service packs or hotfixes.
If you have a server that cannot stay up due to memory
leaks or crashes, then you should investigate the cause
of these issues, be it drivers, utilities or applications.
Once you've isolated the problem to a specific bit of
code or hardware, the appropriate action is to replace
or repair the component, or if replacement is not an
option, schedule the appropriate process to be restarted
periodically -- no need to reboot the entire system just
because of a single app.
TOOLS
There are a few tools that will allow you to calculate
the uptime of an NT system, locally or remotely.
They can be found at the following locations:
• SRVINFO ................ Resource Kit
• UPTIME ................. Resource Kit
• SYSTEMINFO ............. Native Utility (WinXP/2003)
• UPTIME ................. http://www.microsoft.com/ntserver/nts/downloads/management/uptime/
• PSINFO ................. http://www.microsoft.com/technet/sysinternals/utilities/PsTools.mspx
• WINUP .................. http://www.freedownloadscenter.com/Network_and_Internet/UNIX_Ports_and_Commands/WinUp.html
USING MICROSOFT'S UPTIME
The version of UPTIME that Microsoft provides on their
site and in the ResKit can provide information about
uptime over a long period, in addtion to current uptime.
Rather than digging around in the EventLog, you can run
UPTIME with a "/P" parameter and get they history of both
uptime and downtime, plus any STOP error messages.
Run the following command at a CMD prompt:
UPTIME /P:30
This will give your system's uptime over the past 30 days
including any Blue Screens, with the STOP ERROR #
Using the /HEARTBEAT parameter will allow the system to
write a date/time stamp at a fixed interval, so that if
your system freezes, upon reboot, a reasonable timeframe
of the incident can be estimated.
WHITEPAPERS & TECH DOCUMENTS
• http://support.microsoft.com/?KBID=232243
• http://www.computerperformance.co.uk/no_more_reboot.htm
PERSONAL NOTES
• If you're rebooting more than once per quarter as
a rule, you really should consider re-evaluating
your operating procedures, or the software/hardware
that is being deployed.
• SRVINFO also provides other statistics and system
information.
• SYSTEMINFO, available with Windows XP, provides a
superset of the functionality found in SRVINFO.
• For best effect, enable the HEARTBEAT parameter of
Microsoft's UPTIME utility.
