IBM Mainframe Forum

[bazzigar]

Hi,

Is there any way I can get the system down time & IPL time from the Syslog. Any specific message ID's from the Syslog says that?

I know we will get IPL time from D IPLINFO. But from which message ID the D IPLINFO take the time from?
Hope I dont confuse all

Thanks,

[steve-myers]

Shutdown

One would think it would be recorded in SMF, but it isn't!

IPL time

No message. It's stashed in a system control block. 20 years ago, I knew. I still run a program that retrieves and formats it but the source is long gone.

[bazzigar]

So from where/how I can get the System down time? So that I can calulate the total outage.

[steve-myers]

In my experience, z/OS systems tend to be unusable long before they are actually shut down, even when unusable, some subsystems may be active and at least partially usable even when the system as a whole is unusable. You can usually directly retrieve from SMF data when individual subsystems are taken down during a system shut down. Of course, if the system died when the power grid died, no system data may be recorded.

Similarly, IPL time does not reveal when the major subsystems become active; this may not happen for some time after an IPL. Many subsystems require the network to be running before they can be considered usable.

[Robert Sample]

calulate the total outage.

Perhaps you need to sit back and define "total outage" before proceeding? Even if you come up with a system down time (which depends at least partly on what is going on to cause the system to need the IPL, and as Mr. Myers indicates can be tough to determine since different parts of the system may be shut down at different times), how are you going to determine the restart time? The time the IPL starts can be 15 to 45 minutes before a TSO or CICS user can log in, for example. Is the "total outage" just the 20 minutes from system being locked up to the IPL start, or is it the 55 minutes between the system being locked up and the time a CICS or TSO user can actually log in and use the system again? Unless there's a clear definition of "total outage", both numbers are equally valid in this case -- depending upon your definition of "total outage". And it could have been 120 minutes between the time the TSO user last hit the enter key and when the TSO user can resume the edit session and hit the enter key again.

[enrico-sorichetti]

seems silly to me to talk about system downtime and friends...
what is important is the APPLICATION uptime/downtime
the system might be up and happily running,
but if the master customer database is unavailable the above is completely irrelevant

[Anuj Dhawan]

Perhaps you'd like to use IPLINFO from Mark Zelden.

[jaggz]

Hi,

Determing the System Downtime and Start Up time is purely a Shop Specific. Hopefully you should have noticed this while IPLing your Test LPAR. Planned Outage will give you a perfect Picture.(So you can Start Calculating the Time)

HTH

Jaggz