X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C734F8.D78C1B74@onstor-exch02.onstor.net>; Wed, 10 Jan 2007 12:49:37 -0800
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-class: urn:content-classes:message
Subject: RE: time difference issue
Date: Wed, 10 Jan 2007 12:49:38 -0800
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E01FF8D61@onstor-exch02.onstor.net>
In-Reply-To: <BB375AF679D4A34E9CA8DFA650E2B04E028FD1@onstor-exch02.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: time difference issue
thread-index: Acc08rrMJv9eJyvEQvyAHrkcSnVOcgAAtVdwAAArfJAAABTOoAAAdzwg
From: "Maxim Kozlovsky" <maxim.kozlovsky@onstor.com>
To: "Shamsudeen Jeseem" <jeseem@onstor.com>,
	"John Rogers" <john.rogers@onstor.com>,
	"Andy Sharp" <andy.sharp@onstor.com>,
	"Chris Vandever" <chris.vandever@onstor.com>
Cc: "John Keiffer" <john.keiffer@onstor.com>,
	"dl-Software" <dl-software@onstor.com>,
	"dl-QA" <dl-qa@onstor.com>

Well, if the bios battery really needs replacement, there is no reason
to prevent the cluster from functioning when simply running ntpdate at
bootup works around that problem.=20

> -----Original Message-----
> From: Shamsudeen Jeseem=20
> Sent: Wednesday, January 10, 2007 12:37 PM
> To: John Rogers; Andy Sharp; Chris Vandever
> Cc: John Keiffer; dl-Software; dl-QA
> Subject: RE: time difference issue
>=20
> when changing the system time on the filer.
> or when adding a new NTP server, to get the initial time sync.
>=20
> We shouldn't have to do this at each bootup ( if it is=20
> failing at each bootup, it mean bios battery needs replacement)
>=20
>=20
>=20
>=20
> -----Original Message-----
> From: John Rogers
> Sent: Wednesday, January 10, 2007 12:32 PM
> To: Shamsudeen Jeseem; Andy Sharp; Chris Vandever
> Cc: John Keiffer; dl-Software; dl-QA
> Subject: RE: time difference issue
>=20
> When do we execute ntpdate on our system? Is it at boot time=20
> and config create time?
>=20
> -----Original Message-----
> From: Shamsudeen Jeseem
> Sent: Wednesday, January 10, 2007 12:31 PM
> To: Andy Sharp; Chris Vandever
> Cc: John Keiffer; dl-Software; dl-QA
> Subject: RE: time difference issue
>=20
> The NTP daemon threshold is 1000seconds. Beyond this the NTP=20
> daemon will exit.
> The daemon usually takes 2000 seconds, for correcting a skew=20
> of 1 second beyond the acceptable range
>=20
> -jeseem=20
>=20
> -----Original Message-----
> From: Andrew Sharp [mailto:andy.sharp@onstor.com]
> Sent: Wednesday, January 10, 2007 12:06 PM
> To: Chris Vandever
> Cc: John Keiffer; dl-Software; dl-QA
> Subject: Re: time difference issue
>=20
> The ntp daemon itself also has a threshold for the time=20
> difference, beyond which it will give up and not try to close=20
> the gap, as it deems it to be too great.  For cases where the=20
> time delta is too great, the ntpdate program should be used=20
> to jump the date forward (or backward).
> After taking into consideration the things Chris mentions=20
> below, of course.
>=20
> Cheers,
>=20
> a
>=20
>  On Wed, 10 Jan 2007 11:39:11 -0800 "Chris Vandever"
> <chris.vandever@onstor.com> wrote:
>=20
> > I believe it takes awhile for NTP to adjust the time to=20
> what it gets=20
> > from the NTP server.  If we were to simply change it=20
> immediately a lot=20
> > of things might break (premature timeouts, etc.),=20
> especially if we're=20
> > 'fast'.
> >=20
> > =20
> >=20
> > Forwarding to dl-QA, as they have more experience than I do on how=20
> > long you should expect it to take for NTP to synchronize=20
> time across=20
> > multiple nodes.
> >=20
> > =20
> >=20
> > ChrisV
> >=20
> > =20
> >=20
> > ________________________________
> >=20
> > From: John Keiffer
> > Sent: Wednesday, January 10, 2007 11:30 AM
> > To: dl-Software
> > Subject: FW: time difference issue
> >=20
> > =20
> >=20
> > Hello all,
> >=20
> > =20
> >=20
> > Please see my original email below. Can someone explain further?
> >=20
> > =20
> >=20
> > Thanks,
> >=20
> > John
> >=20
> > =20
> >=20
> > ________________________________
> >=20
> > From: Nagendra Subramanya
> > Sent: Wednesday, January 10, 2007 11:19 AM
> > To: John Keiffer
> > Subject: RE: time difference issue
> >=20
> > =20
> >=20
> > John,
> >=20
> > I don't know about other daemons that are affected by=20
> addition/removal=20
> > of NTP servers.
> >=20
> > I think clusterDB related daemons may be affected.
> >=20
> > Please post your question to dl-Software so that others may=20
> comment on=20
> > your scenario.
> >=20
> > =20
> >=20
> > -Nags
> >=20
> > =20
> >=20
> > ________________________________
> >=20
> > From: John Keiffer
> > Sent: Wednesday, January 10, 2007 9:05 AM
> > To: Nagendra Subramanya
> > Subject: time difference issue
> >=20
> > =20
> >=20
> > Hey Nagendra,
> >=20
> > =20
> >=20
> > I know you said the following message is just=20
> informational, but I'm=20
> > trying to understand in which places the time is computed to be=20
> > different. I added NTP servers to both my nodes, but found that the=20
> > ssc_console times where still not correct. Is this where the time=20
> > difference is happening? After I rebooted BOTH nodes, then=20
> everything=20
> > was in sync and NCM was accessible again.
> >=20
> > =20
> >=20
> > "auth_agent:INFO: authAgent_checkForReqTimeouts: 5825:=20
> Timedifference=20
> > is computed to be greater than 1 sec: 2"
> >=20
> > =20
> >=20
> > Thank you,
> >=20
> > John Keiffer
> >=20
> > =20
> >=20
> > =20
> >=20
>=20
