Thanks for all who replied. I will try some of them
later and update again.
*******************************************************
alan_at_nabeth.cxo.dec.com <alan_at_nabeth.cxo.dec.com>wrote
Since the message comes from the savecore, I looked at
to get the context. It seems to be printed when the
savecore program exits with a status 3. The manual
page doesn't describe what the different status codes
mean. You can either run the savecore script or
savecore itself to see what messages it gives.
My guess is that at the time savecore runs one of the
dump spaces isn't available. It could be a timing
interaction with LSM.
1. I'd guess not, if savecore is complaining of the
raw
dump device space.
2. Since it comes from running the savecore startup
script, probably not.
3. Depends on what's wrong.
4. See previous guess.
********************************************-----
Wolfers, Ray <wolfers_at_rlmsystems.com.au> wrote
LSM introduces a problem when processing crashdumps.
> >
> > By default LSM will recover volumes that are
mirrored. I assume that both your swap disks are
mirrored. When LSM starts to recover your swap
volumes, it destroys any crashdump that may be there.
Note that this is only the case ,if your swap volumes
are used, ie. if your machine is paging to the swap
disks. Therefor you will never be able to get savecore
to do a crashdump analysis.
> >
> > There is a work-around. Set the "norecov" option
for your swap volumes
> only.
> > This is recommended in the LSM manual.
> >
> > Start dxlsm
> > Select the swap volume so that it is highlighted
> > Right-mouse click on the swap volume to get its
properties
> > Change "Startup:" option to "norecov". Apply .
> > Repeat for your other swap volumes
> >
> > On startup LSM will not recover the swap volumes
and savecore will now be able to process crashdummps.
You will no longer get that error message again.
++++++++++++++++++++++++++++++++++++++++++++++
Ray Wolfers
> r.wolfers_at_rlmsystems.com.au wrote again.
> It should not matter what the usage type is, as long
as you have set norecov on a volume. Then when LSM
attempts to recover all volume mirrors after a crash
(or when manually run) it will not resync those for
which startup=norecov.
>
> crashdump header will be written to swap1 whilst the
crashdump is written on swap2 (continues onto swap1 if
crashdump is larger than swap2). So you must set
startup=norecov for both swap1 and swap2.
>
> Works for us!
>
> You will not see the norecov option described in any
LSM output. I think the only place where you can see
and set it is in dxlsm.
>
> You should notice that you get no resynchronisation
messages for your swap volumes (from volwatch) as your
machine recovers its volumes.
>
> Good luck
>
*****************************************> >
> aelee_at_household.com <aelee_at_household.com> wrote
I think this message is informational only. I also
see it even when doing a
routine
reboot.
Annette Lee
Sysadm, Household Intl
***************************************************
Dejan Muhamedagic <dejan_at_quant-x.com> wrote
Hello,
> we do have some old crash files/vmcore files exist
in
> $DIR. we run 4.0D PK3.
>
> my doubt is:
>
> (1)if I delete those old crash files, will this
error
> go?
no.
> (2)is there any place this error message is saved. i
> can't see in messages/kern.log etc.
> (i could see a similarlines in savecore script which
> is validating a case statement.)
don't think so but you may try /var/adm/messages.
> (3)is there any thing to be changed in the system?
no.
> (4)what does the system trying to tell me?
that there's no dump to be saved, so you shouldn't
worry about this.
Bye.
*******************************************************
My original question
> -----Original Message-----
> From: uunniixx_aaddmmin_at_yahoo.com
[mailto:uunniixx_aaddmmin_at_yahoo.com]
> Sent: Friday, October 08, 1999 10:56 PM
> Subject:
>
> >
> > -----Original Message-----
> > From: unix adm
[mailto:uunniixx_aaddmmin_at_yahoo.com]
> > Sent: Friday, October 08, 1999 8:01 AM
> > To: Tru64-Unix Managers
> > Subject: message while booting
> >
> >
> > Dear managers
> >
> > our system gives a message "some or all of the
> > partitions which might contain a dump header could
not
> > be searched" when booting.I can't catch another
> > message associated withthat.it was fast.(may be
"no
> > dump was found". I am guessing by reading the
savecore
> > script).
> > I beleive this is shown while system executes
> > /init.d/savecore script and checks for some crash
> > dumps.
> >
> > we have 2 swap disks, both running under LSM.one
is a
> > dedicated disk and another is a partition.
> >
> > we do have some old crash files/vmcore files exist
in
> > $DIR. we run 4.0D PK3.
> >
> > my doubt is:
> >
> > (1)if I delete those old crash files, will this
error
> > go?
> > (2)is there any place this error message is saved.
i
> > can't see in messages/kern.log etc.
> > (i could see a similarlines in savecore script
which
> > is validating a case statement.)
> > (3)is there any thing to be changed in the system?
> > (4)what does the system trying to tell me?
> >
> > any help will save me lot of time and may stop
some
> > more hair falling.
> >
> > Thanks in Advance
> > Duadmin
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Bid and sell for free at http://auctions.yahoo.com
>
>
>
>
> __________________________________________________
>
> Do You Yahoo!?
>
> Bid and sell for free at http://auctions.yahoo.com
__________________________________________________
Do You Yahoo!?
Bid and sell for free at
http://auctions.yahoo.com
Received on Mon Oct 11 1999 - 14:49:03 NZDT