sources.redhat.com downtime

sources.redhat.com downtime

[Christopher Faylor]

I have been negotiating with various parties on downtime for
sources.redhat.com so that I can make the switch to the new system.

So far, it looks like 1PM on Saturday is doable.  Pessimistically
speaking, I assume that the system will probably be down until
Sunday morning.

I'd hoped to provide more advance notice than this but things didn't
work out that way.

If there are problems, let me know and I'll reschedule.  I may have
to do that anyway but it was getting close enough to Saturday now that
I thought I should let people know even though I don't have a 100%
firm date/time yet.

cgf

Re: sources.redhat.com downtime

[Andrew Cagney]

> I have been negotiating with various parties on downtime for
> sources.redhat.com so that I can make the switch to the new system.
> 
> So far, it looks like 1PM on Saturday is doable.  Pessimistically
> speaking, I assume that the system will probably be down until
> Sunday morning.

Who's 1pm saturday?

Andrew


> I'd hoped to provide more advance notice than this but things didn't
> work out that way.
> 
> If there are problems, let me know and I'll reschedule.  I may have
> to do that anyway but it was getting close enough to Saturday now that
> I thought I should let people know even though I don't have a 100%
> firm date/time yet.
> 
> cgf
> 

Re: sources.redhat.com downtime

[Benjamin Kosnik]

Right on! Thanks Christopher for your efforts and the advance notice.

-benjamin

Re: sources.redhat.com downtime

[Christopher Faylor]

On Thu, Jan 16, 2003 at 04:21:19PM -0500, Andrew Cagney wrote:
>>I have been negotiating with various parties on downtime for
>>sources.redhat.com so that I can make the switch to the new system.
>>
>>So far, it looks like 1PM on Saturday is doable.  Pessimistically
>>speaking, I assume that the system will probably be down until
>>Sunday morning.
>
>Who's 1pm saturday?

Wow, this was predictable.

Put it this way.  If you try to access sources.redhat.com and find that
it is unavailable then that will be a good indication of which I meant.

cgf

Re: sources.redhat.com downtime

[Jonathan Larmour]

Christopher Faylor wrote:
> I have been negotiating with various parties on downtime for
> sources.redhat.com so that I can make the switch to the new system.
> 
> So far, it looks like 1PM on Saturday is doable.  Pessimistically
> speaking, I assume that the system will probably be down until
> Sunday morning.

1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
since presumably you're going to be physically there.

Jifl
-- 
eCosCentric       http://www.eCosCentric.com/       <info@eCosCentric.com>
--[ "You can complain because roses have thorns, or you ]--
--[  can rejoice because thorns have roses." -Lincoln   ]-- Opinions==mine

Confirmed: sources.redhat.com downtime

[Christopher Faylor]

On Thu, Jan 16, 2003 at 10:24:24PM +0000, Jonathan Larmour wrote:
>Christopher Faylor wrote:
>>I have been negotiating with various parties on downtime for
>>sources.redhat.com so that I can make the switch to the new system.
>>
>>So far, it looks like 1PM on Saturday is doable.  Pessimistically
>>speaking, I assume that the system will probably be down until
>>Sunday morning.
>
>1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
>since presumably you're going to be physically there.

The timezone in my Date: field accurately depicts my point of reference.
I'm not going to be physically anywhere near either machine.

Unless someone says that this isn't enough advance notice, I've finally
confirmed that my proposed downtime is doable.  The downtime will be:

    Thu 2003-01-16 18:00 GMT

    Thu 2003-01-16 10:00 US/Pacific
    Thu 2003-01-16 11:00 US/Arizona
    Thu 2003-01-16 11:00 US/Mountain
    Thu 2003-01-16 12:00 US/Central
    Thu 2003-01-16 13:00 US/Eastern
    Thu 2003-01-16 13:00 Canada/Eastern
    Thu 2003-01-16 16:00 America/Sao_Paulo
    Thu 2003-01-16 18:00 Europe/London
    Thu 2003-01-16 19:00 Europe/Berlin
    Thu 2003-01-16 13:00 Australia/Victoria
    Thu 2003-01-16 13:00 Australia/Sydney

The system will be down until it comes back up.  It will be three hours
at a minimum.

cgf

Re: sources.redhat.com downtime

[Joseph S. Myers]

On Thu, 16 Jan 2003, Jonathan Larmour wrote:

> 1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
> since presumably you're going to be physically there.

Could the new system use a consistent timezone (UTC)?  At present,
messages generated by CVS and GNATS do get sent out with times in UTC
(albeit specified as "-0000" which can also mean unknown timezone rather
than "+0000" as recommended in RFC2822), but the web list archives display
times in a local timezone.  (Timezone changes would require adjustments to
cron jobs, since those are in local time and adjusted to levels of load on
the machine.)

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: Confirmed: sources.redhat.com downtime

[Andrew Cagney]

> Unless someone says that this isn't enough advance notice, I've finally
> confirmed that my proposed downtime is doable.  The downtime will be:

Oh, no it won't!  :-)

But it will be at these times +48:00 :-)

>     Thu 2003-01-16 18:00 GMT
> 
>     Thu 2003-01-16 10:00 US/Pacific
>     Thu 2003-01-16 11:00 US/Arizona
>     Thu 2003-01-16 11:00 US/Mountain
>     Thu 2003-01-16 12:00 US/Central
>     Thu 2003-01-16 13:00 US/Eastern
>     Thu 2003-01-16 13:00 Canada/Eastern
>     Thu 2003-01-16 16:00 America/Sao_Paulo
>     Thu 2003-01-16 18:00 Europe/London
>     Thu 2003-01-16 19:00 Europe/Berlin

But I don't know what happened here

>     Thu 2003-01-16 13:00 Australia/Victoria
>     Thu 2003-01-16 13:00 Australia/Sydney
> 
> The system will be down until it comes back up.  It will be three hours
> at a minimum.

Just sending something to the gdb list.

Andrew

Re: Confirmed: sources.redhat.com downtime

[Christopher Faylor]

On Thu, Jan 16, 2003 at 05:40:49PM -0500, Christopher Faylor wrote:
>On Thu, Jan 16, 2003 at 10:24:24PM +0000, Jonathan Larmour wrote:
>>Christopher Faylor wrote:
>>>I have been negotiating with various parties on downtime for
>>>sources.redhat.com so that I can make the switch to the new system.
>>>
>>>So far, it looks like 1PM on Saturday is doable.  Pessimistically
>>>speaking, I assume that the system will probably be down until
>>>Sunday morning.
>>
>>1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
>>since presumably you're going to be physically there.
>
>The timezone in my Date: field accurately depicts my point of reference.
>I'm not going to be physically anywhere near either machine.
>
>Unless someone says that this isn't enough advance notice, I've finally
>confirmed that my proposed downtime is doable.  The downtime will be:
>
>    Thu 2003-01-16 18:00 GMT

Sorry.

      Sat 2003-01-18 18:00 GMT

      Sat 2003-01-18 10:00 US/Pacific
      Sat 2003-01-18 11:00 US/Arizona
      Sat 2003-01-18 11:00 US/Mountain
      Sat 2003-01-18 12:00 US/Central
      Sat 2003-01-18 13:00 US/Eastern

cgf

Re: sources.redhat.com downtime

[Christopher Faylor]

On Thu, Jan 16, 2003 at 10:40:29PM +0000, Joseph S. Myers wrote:
>On Thu, 16 Jan 2003, Jonathan Larmour wrote:
>
>> 1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
>> since presumably you're going to be physically there.
>
>Could the new system use a consistent timezone (UTC)?  At present,
>messages generated by CVS and GNATS do get sent out with times in UTC
>(albeit specified as "-0000" which can also mean unknown timezone rather
>than "+0000" as recommended in RFC2822), but the web list archives display
>times in a local timezone.  (Timezone changes would require adjustments to
>cron jobs, since those are in local time and adjusted to levels of load on
>the machine.)

I can set the time to GMT on the new system, if everybody agrees that is
reasonable.  It seems reasonable to me.

cgf

Re: sources.redhat.com downtime

[Andrew Cagney]

> On Thu, Jan 16, 2003 at 10:40:29PM +0000, Joseph S. Myers wrote:
> 
>>On Thu, 16 Jan 2003, Jonathan Larmour wrote:
>>
> 
>>> 1pm your time, 1pm sources.redhat.com's time or what? :-) I guess PST 
>>> since presumably you're going to be physically there.
> 
>>
>>Could the new system use a consistent timezone (UTC)?  At present,
>>messages generated by CVS and GNATS do get sent out with times in UTC
>>(albeit specified as "-0000" which can also mean unknown timezone rather
>>than "+0000" as recommended in RFC2822), but the web list archives display
>>times in a local timezone.  (Timezone changes would require adjustments to
>>cron jobs, since those are in local time and adjusted to levels of load on
>>the machine.)
> 
> 
> I can set the time to GMT on the new system, if everybody agrees that is
> reasonable.  It seems reasonable to me.

Yes please.  This was discussed before, the consensus then being that a 
new machine could have this change (just not the old machine).

Andrew

Re: sources.redhat.com downtime

[Jason Molenda]

On Thu, Jan 16, 2003 at 06:13:18PM -0500, Andrew Cagney wrote:

> > I can set the time to GMT on the new system, if everybody agrees that is
> > reasonable.  It seems reasonable to me.
> 
> Yes please.  This was discussed before, the consensus then being that a 
> new machine could have this change (just not the old machine).

Fine by me.  Originally I put everything (including the mhonarc
archives) in US/Pacific because that's where I was :-), but it's
a little tricky to justify its continued use...

J (I think Joseph wants UTC just because he lives so close :)

Re: sources.redhat.com downtime

[Joseph S. Myers]

On Thu, 16 Jan 2003, Andrew Cagney wrote:

> > I can set the time to GMT on the new system, if everybody agrees that is
> > reasonable.  It seems reasonable to me.
> 
> Yes please.  This was discussed before, the consensus then being that a 
> new machine could have this change (just not the old machine).

Thanks, just note if fixing up crontabs that at least gccadmin's crontab
is CVS-controlled (/cvs/gcc/gcc/maintainer-scripts/) - I don't know about
others, but presume in general that jobs should continue to run at the
same absolute time each day.

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: sources.redhat.com downtime

[Christopher Faylor]

On Fri, Jan 17, 2003 at 11:19:06AM +0000, Joseph S. Myers wrote:
>On Thu, 16 Jan 2003, Andrew Cagney wrote:
>
>> > I can set the time to GMT on the new system, if everybody agrees that is
>> > reasonable.  It seems reasonable to me.
>> 
>> Yes please.  This was discussed before, the consensus then being that a 
>> new machine could have this change (just not the old machine).
>
>Thanks, just note if fixing up crontabs that at least gccadmin's crontab
>is CVS-controlled (/cvs/gcc/gcc/maintainer-scripts/) - I don't know about
>others, but presume in general that jobs should continue to run at the
>same absolute time each day.

Is there any reason why I can't leave that to the maintainers of the individual
crontab owners?  I do have all of the crontab entries running, but getting the
times right could be tricky.  You can't just add 8 to the hours, obviously.

Does anyone have a "move crontab to a new timezone" script, maybe?

cgf

Re: sources.redhat.com downtime

[Alexandre Oliva]

On Jan 17, 2003, Christopher Faylor <cgf@redhat.com> wrote:

> Does anyone have a "move crontab to a new timezone" script, maybe?

There's always the option of running crond with the Pacific timezone,
to give maintainers more time to prepare to the switch-over.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                 aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp        oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist                Professional serial bug killer

Re: sources.redhat.com downtime

[Christopher Faylor]

On Fri, Jan 17, 2003 at 01:59:19PM -0200, Alexandre Oliva wrote:
>On Jan 17, 2003, Christopher Faylor <cgf@redhat.com> wrote:
>> Does anyone have a "move crontab to a new timezone" script, maybe?
>
>There's always the option of running crond with the Pacific timezone,
>to give maintainers more time to prepare to the switch-over.

I thought of that but that sort of eliminates the whole reason for the
switch.

I'll probably have lots of time during tomorrow's rsync to adjust
everything if no one can point me to an automated way of doing this.  I
can even make the new system available during the switch if people want
to logon and fix their crontab entries.

cgf

Re: sources.redhat.com downtime

[Andrew Cagney]

> On Fri, Jan 17, 2003 at 11:19:06AM +0000, Joseph S. Myers wrote:
> 
>>On Thu, 16 Jan 2003, Andrew Cagney wrote:
>>
> 
>>> > I can set the time to GMT on the new system, if everybody agrees that is
>>> > reasonable.  It seems reasonable to me.
> 
>>> 
>>> Yes please.  This was discussed before, the consensus then being that a 
>>> new machine could have this change (just not the old machine).
> 
>>
>>Thanks, just note if fixing up crontabs that at least gccadmin's crontab
>>is CVS-controlled (/cvs/gcc/gcc/maintainer-scripts/) - I don't know about
>>others, but presume in general that jobs should continue to run at the
>>same absolute time each day.
> 
> 
> Is there any reason why I can't leave that to the maintainers of the individual
> crontab owners?  I do have all of the crontab entries running, but getting the
> times right could be tricky.  You can't just add 8 to the hours, obviously.

None.

In fact I think gcc and gdb's cron should start out disabled.  Relevent 
admins can re-start them with adjusted timezones once other things have 
settled.

As for a script, don't bother.

Andrew

Re: sources.redhat.com downtime

[Christopher Faylor]

On Fri, Jan 17, 2003 at 11:12:36AM -0500, Andrew Cagney wrote:
>>On Fri, Jan 17, 2003 at 11:19:06AM +0000, Joseph S. Myers wrote:
>>
>>>On Thu, 16 Jan 2003, Andrew Cagney wrote:
>>>
>>
>>>>> I can set the time to GMT on the new system, if everybody agrees that 
>>>>is
>>>>> reasonable.  It seems reasonable to me.
>>
>>>>
>>>>Yes please.  This was discussed before, the consensus then being that a 
>>>>new machine could have this change (just not the old machine).
>>
>>>
>>>Thanks, just note if fixing up crontabs that at least gccadmin's crontab
>>>is CVS-controlled (/cvs/gcc/gcc/maintainer-scripts/) - I don't know about
>>>others, but presume in general that jobs should continue to run at the
>>>same absolute time each day.
>>
>>
>>Is there any reason why I can't leave that to the maintainers of the 
>>individual
>>crontab owners?  I do have all of the crontab entries running, but getting 
>>the
>>times right could be tricky.  You can't just add 8 to the hours, obviously.
>
>None.
>
>In fact I think gcc and gdb's cron should start out disabled.  Relevent 
>admins can re-start them with adjusted timezones once other things have 
>settled.

That is actually what I've done.  The crontab entries are there but they
are commented out.

cgf
(already stressed with a day to go)

Re: sources.redhat.com downtime

[Hans-Peter Nilsson]

On Fri, 17 Jan 2003, Christopher Faylor wrote:
> (already stressed with a day to go)

Deep breaths.  Don't panic.  Almost there.  Steady as she goes.
(htdig crontab entries trigging the next days now commented
out.)

brgds, H-P