Daylight Saving Time patch

[ScottLClement]

Does anyone know the patch name to fix the daylight saving time switch

[Denver Osborn]

For hpux 11i v1 PHCO_34668, PHCO_34673 for 11.0

-denver

[James R. Ferguson]

Hi Scott:

The patch you are looking for (as noted) is for the '/usr/lib/tztab' file:

For 11.0 PHCO_34673
For 11.11 PHCO_34668
For 11.23 PHCO_34669

The patch does not cause a reboot when applied. However, after the update of the 'tztab' file, long-running processes (notably 'cron' and any databases) should be restarted. A a reboot sometime before the timezone(s)' rule changes will mean that any code using, standard C library system calls, will reread and recache the updated rules.

Regards!

...JRF...

[Ray Ball]

I don't have the technical expertise to contradict the above advice regarding the need to reboot ALL HP-UX servers throughout an organization (which can be rather disruptive), so I sought another opinion from HP's Response Center. Unfortunately, I got the opposite answer, with one exception -- Java. See: http://www.hp.com/products1/unix/java/DST-US.html

I took great pains to pose the question so as to distinguish it from the obvious answer that "installing the patch" does not cause a reboot versus the downstream effect on already running processes like cron. So I have to put a certain amount of faith in HP because their response seemed to address my concern about the effect being dynamic and not requiring processes to be stopped & restarted.

Now I don't know who to put my faith in. The safest approach is to disrupt the corporation and reboot all servers. We have some that haven't been rebooted in over 1000 days and don't look forward to creating unnecessary disruption without some assurance that it was ABSOLUTELY necessary.

Frankly, it strikes me that a change of this magnitude, i.e., causing a shop to reboot all of its HP-UX servers, ought to induce HP to make a definitive statement aimed directly at their HP-UX customer base and put the question to rest for everyone - once and for all. In fact, that was the approach I took when I worded my question to the Response Center. But since they seem to think that there is no need to reboot to cause the desired effect to take place, I guess they see no need for such a broadcast.

Intuitively I side with the advice to reboot, but selling a major disruption to management based on a forum comment that doesn't agree with the response directly from HP presents a problem for me.

Where can I get an authoriative tie-breaker ;-)

Thanks...Ray

[Pete Randall]

Ray,

If you read JRF's answer carefully, he says "long-running processes (notably 'cron' and any databases) should be restarted". Restarted and rebooted are two very different things. Yes, Jim goes on and suggests a reboot just to make sure that EVERYTHING that should be restarted gets restarted, but with sufficient knowledge of your system and applications, this can be avoided by restarting just those things that require it, like cron and any databases.


Pete

[Ray Ball]

Pete:

I'm not trying to be agrumentative, but requiring every customer to rely on "sufficient knowledge" of long running processes and applications is highly problematical. Especially in an age of purchased applications. In my narrow mind, the ONLY way to CAUSE the desired effect to take place is to restart ALL processes and applications, which is tatamount to rebooting. Frankly, my argument is with HP because I fear that many customers will see the "Automatic Reboot? No" statement in the patch and go on their merry way, only to get clobbered when long-running processes, i.e., cron, and databases fail to act properly come March 11th. I think this warrants a more proactive approach from HP. Does this sound unreasonable?

Ray

[Dennis Handly]

>Ray: In my narrow mind, the ONLY way to CAUSE the desired effect to take place is to restart ALL processes and applications, which is tatamount to rebooting.

Except you can do it at your leisure and stopping applications separately may be easier than rebooting.

>I think this warrants a more proactive approach from HP. Does this sound unreasonable?

It isn't unreasonable but what can HP say if they didn't write the application??

And suppose the application never uses localtime(3) or ctime(3)?