Re: 8.1.7 install problems on Sun Blade 100/Solaris 8 01/01 MU4

"Howard J. Rogers" wrote:
>
> I of course didn't notice that you are trying to get to the nomount stage in
> order to create a database in the first place. My apologies. Forget that
> stuff about checking permissions on your controlfiles, data files and redo
> logs, since they won't be there yet!! Do still check the pfile itself, of
> course.
>

My permissions on the pfile look fine. I'm starting to think there is something incompatible with the Solaris patches on the Blade and Oracle 8.1.7. I have no problem installing 8.1.7 on a Solaris 2.6 E450.

Thanks, Greg

>
> "Howard J. Rogers" <howardjr_at_www.com> wrote in message
> news:3b10e429$1_at_news.iprimus.com.au...
> > Out of left field, and I'm sure you Unix types wouldn't have allowed this
 to
> > happen, but I regularly generate 'end of file on communication channel'
> > errors on training courses by setting the permissions on the relevant
 files
> > incorrectly, so that the Oracle processes don't have rights to use them.
 I
> > of course fix this with a quick chmod 777 on the files in question, but
 Unix
> > users usually laugh at this point, so I guess there is a more intelligent
> > choice!!
> >
> > Since you get nothing after specifying the pfile, I'd check the ownership
> > and privileges on that. I'd also have a look at all your data files, redo
> > logs and controlfiles whilst you're at it.
> >
> > The trouble is, of course, that ORA-3113 covers a host of sins, and
 there's
> > no telling that this is (or isn't) one of them until you've checked.
> >
> > Regards
> > HJR
> > --
> > =============================!!=============================
> > The views expressed are my own only, and definitely NOT those of Oracle
> > Corporation
> > =============================!!=============================
> >
> >
> > "G.T." <ethant_at_pacificnet.net> wrote in message
> > news:3B0FFB64.17D959F0_at_pacificnet.net...
> > > Sybrand Bakker wrote:
> > > >
> > > > >
> > > > > Yes, I've checked the above, I've rebooted, and my environment has
> > > > > ORACLE_SID, etc.
> > > > >
> > > > > Here is what happens when I run svrmgr:
> > > > >
> > > > > SVRMGR> connect internal
> > > > > Connected.
> > > > > SVRMGR> startup pfile =
> > > > > "/export/home0/u01/app/oracle/admin/records/pfile/initrecords.ora"
> > > > > ORA-03113: end-of-file on communication channel
> > > > >
> > > > > I found one item on the web regarding Solaris 8 and Oracle 8.1.6
 where
> > > > > they advise to add a bunch of patches from Oracle but I assume those
> > > > > patches were already rolled into 8.1.7.
> > > > >
> > > >
> > > > Usually the SID is limited to 4 characters.
> > >
> > > I'm only familiar with 7.3 and up and all our SIDs are 6 or 7
> > > characters.
> > >
> > > > You also should verify you alert log in the directory pointe d to by
 the
> > > > background_dump_dest parameter.
> > > > In a typical install this is the
> > > > /export/home0/u01/app/oracle/admin/records/bdump
> > > >
> > >
> > > That directory exists but there is no alert log there.
> > >
> > > > One of the causes for 3113 is incorrect init.ora parameter, which
 results in
> > > > memory problems, which results in the instance not starting.
> > > >
> > >
> > > I've got 384MB RAM and looking through my init.ora everything is set to
> > > a small implementation. Anything in particular I should look for?
> > >
> > > TIA, Greg
> > >
> > > --
> > > See http://www.2fortheroad.net for ugly bikes and stale music.
> > > "Why does a man with nothing to say, he always seems to talk?" - Marc
> > > Riley
> >
> >
 

-- 
See http://www.2fortheroad.net for ugly bikes and stale music. 
"Why does a man with nothing to say, he always seems to talk?" - Marc
Riley