Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: dbwr and lgwr don't shutdown

Re: dbwr and lgwr don't shutdown

From: NetComrade <netcomradeNSPAM_at_bookexchange.net>
Date: Thu, 31 Mar 2005 22:44:52 GMT
Message-ID: <424c6e4a.175315330@localhost>


I apologize.. Let me re-state the problem

  1. v880 and 2 e4500 (nodes) are connected ot same storage via fibre hubs in Arbitraded Loops, so this is an old SAN
  2. each node controls different disks via Veritas Volume Manager
  3. each node is capable of running each other's databases. Databases/disks are grouped, and groups can be swung over by Veritas Cluster server
  4. e4500 run solaris 7
  5. v880 is running solaris 8 (it can't run 7)
  6. the file system is vxfs

This is a 'cluster', but there is always a once instance of a database. The only 'cluster' thing about it, is that instances can fail over if node/host fails.

e4500 has no problem opening/closing the TEST database (oracle_sid=TEST). The V880 has problems.. but didn't have problems earlier.. this happenned after we upgraded OBP on the server.. but the machine had too many issues prior to that, so we are currently rebuilding the box.

On Thu, 31 Mar 2005 20:20:13 GMT, HansF <News.Hans_at_telus.net> wrote:

>On Thu, 31 Mar 2005 19:49:33 +0000, NetComrade wrote:
>
>> the binaries live on the disk array together with the data
>> the disk array is a5200.
>>
>> apologies for not making it clear
>>
>> I tried trussing either processes, they just didn't seem to be doing
>> anything (perhaps I didn't wait long enough though..)
>>
>
>1) Are you saying this is NFS and not SAN or NAS? That leads to a
>possible area of interest.
>
>2) If you insist on top-posting, I will not respond further. It's not
>worth MY time to flip between top and bottom, and I use bottom (as do
>many, if not most, traditional newsgroup posters).
>
>Top posting is a traditionally accepted method for email, and embedded
>bottom posting (with trim of unneeded info) is the accepted method for
>newsgroups.
>
>Reason - one to one communication in email assumes remembering previous
>communication, and therefore scrolling down is just used as a memory jog.
>Newsgroup posting assumes that anyone in the world will jump in as
>appropriate without prior interaction, so they should be able to follow
>the thread segment from top to bottom where the latest info resides.
>
>/Hans

.......
We use Oracle 8.1.7.4 and 9.2.0.5 on Solaris 2.7 boxes remove NSPAM to email Received on Thu Mar 31 2005 - 16:44:52 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US