Creation/modification time wrong on nfs disk

[Hans Adriaanse]

We have an NFS disk mounted on our OpenVMS system towards a Linux (Redhat) system. On both systems the time is correct (including the timezone). But when I create a file on this NFS disk it is always in the future (and not just one hour). Example:
$ SHOW TIME
6-MAR-2006 08:00:35
$ CREATE TEST.DAT
Exit (Ctrl/Z)
$ DIR/DATE=ALL TEST.DAT

Directory HOME:[HADRIAAN.CLS.SRC]

TEST.DAT;1 6-MAR-2006 09:11:53.00 6-MAR-2006 09:11:53.00 ecified> ecified>

Total of 1 file.
$

The file is 1 hour and 11 minutes in the future ?????

Has anybody an idea and/or suggestions to correct this. MMS does not like this kind of game.

Thanx in advance,
Hans Adriaanse

[Wim Van den Wyngaert]

Conmpare your timezone settings. May be they are not the same. The 11 minutes must be found somewhere else (wrong clock ?).

Wim

[Wim Van den Wyngaert]

Compare your timezone settings. May be they are not the same. The 11 minutes must be found somewhere else (wrong clock ?).

Wim

[Wim Van den Wyngaert]

I would say this only once if the interface worked a little better. I used cancel when the blue bar started in IE but it still arrived.

Wim

[Volker Halle]

Hans,

what time does the Linux system show on those files ? Is the TDF (Time Differential Factor) correct on both sides (on VMS: SHOW LOG *TIME*) ?

If you create a file in the NFS-served directory from the Linux system, what time do both systems show ?

Volker.

[Hans Adriaanse]

Yep, the 11 minutes are explained. The linux system had lost its timeserver. That is solved now. But the one hour difference stays.
What is the difference between CET en MET and what should it be in the Netherlands.

[Wim Van den Wyngaert]

MET.

Full spec is :
MET-1MET DST-2,M3.4.0/02,M10.5.0/03
(value of logical sys$timezone_rule)

En sys$timezone_differential should currently be on 3600.

Wim

[Hans Adriaanse]

On one machine this problem occurs. The difference I found with the correct working machines is the version of TCPIP. On the correct working machines it is 5.4 on the problem machine it is 5.3 ECO 2. Is there a known (time)bug in nfs for TCPIP 5.3 ECO 2 ??

[Wim Van den Wyngaert]

I have 7.3 with tcp 5.3 eco 2 as NFS server.

I seem to have the same problem (1 hour difference) on a random number of nodes (7.3 + 5.3/2, 6.2 1h3 + 4.2/1).

Already compared logicals. No differences.

The strange thing is that 1 station created as a clone has the problem while another clone doesn't have the problem.

I continue searching why ...

Wim

[Wim Van den Wyngaert]

I rebooted a station and the problem is solved. It was up for 150 days.

Not yet a clue when/why.

Wim

[Wim Van den Wyngaert · Accepted solution]

Dismounted and mounted (the NFS disk) again. The problem is solved. Can you test that too ?

Wim

[Hans Adriaanse]

Oke, I will try it, but not now. The other users will not like that :-).
Hope to find an occasion soon.
Thanx.

[Wim Van den Wyngaert]

http://www5.itrc.hp.com/service/patch/patchDetail.do?BC=patch.breadcrumb.main|patch.breadcrumb.search|&patchid=DEC-AXPVMS-TCPIP_ECO-V0503-184-4&context=openvms:alpha:7.3

RCP/PRESERVE (aka RCP -p) didn't preserve correct modification timestamp for files created in the summertime (daylight savings
in effect) but copied in the wintertime (standard time), or vice-versa. The timestamp is off by an hour (the daylight savings differential). Problem occurred for files copied to a VMS host, but not for files copied to non-VMS host.

May be this kind of bug is present in NFS ?

Wim

[Wim Van den Wyngaert]

A 6.2 client is persistently wrong, even after a reboot.

Wim

[Hans Adriaanse]

Yes !! A dismount an mount again from the nfs disk fixed the problem. Strange. Thank you all.

[Wim Van den Wyngaert]

But will it stay away ?

I had several systems with the problem. I'll continue to search and report when I find the WHY.

Wim

[Volker Halle]

Hans, Wim,

as far as I remember, you have to re-mount the NFS disk from the client after any changes to the system timezone...

You'll have a chance to verify and test this in 3 weeks ;-)

Volker.

[Wim Van den Wyngaert]

Volker,

No timezone changes.

But I checked all system to see if there is a relation with "time has changed since mount". And this seems to be the case. All 7.3 systems that mounted the disk before the time change have the problem. For the (10) 6.2 systems it is not the case. They are ALWAYS wrong.

Wim