One Instance of two instance/nodes Crash in RAC

Dear all,
I setup two nodes RAC 9.2.0.2 in HP Alpha OpenVMS 8.2 , one node is CAPL21(instance number=1) and another is CAPL22(instance number=2) . when normally run about five days. the second instance CAPL22 terminated abnormally, I check the alert and trace file as follows:
6-AUG-2008 06:29:39.82:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMS1_007.TRC;:
ORA-00600: internal error code, arguments: [kjctr_rksxp:rcvr], [1], [2], [], [], [], [], []
6-AUG-2008 06:29:40.32:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMS1_007.TRC;:
ORA-00603: ORACLE server session terminated by fatal error
ORA-00600: internal error code, arguments: [kjctr_rksxp:rcvr], [1], [2], [], [], [], [], []
6-AUG-2008 06:29:46.60:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMS1_007.TRC;:
ORA-00600: internal error code, arguments: [ksxpcncl3], [], [], [], [], [], [], []
ORA-00603: ORACLE server session terminated by fatal error
ORA-00600: internal error code, arguments: [kjctr_rksxp:rcvr], [1], [2], [], [], [], [], []
6-AUG-2008 06:29:47.38:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMS1_007.TRC;:
ORA-00600: internal error code, arguments: [ksxpcncl3], [], [], [], [], [], [], []
ORA-00603: ORACLE server session terminated by fatal error
ORA-00600: internal error code, arguments: [kjctr_rksxp:rcvr], [1], [2], [], [], [], [], []
6-AUG-2008 06:30:37.33:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_PMON_002.TRC;:
ORA-00484: LMS* process terminated with error
6-AUG-2008 06:30:37.33:
PMON: terminating instance due to error 484
6-AUG-2008 06:30:37.37:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMD0_005.TRC;:
ORA-00484: LMS* process terminated with error
6-AUG-2008 06:30:37.37:
Errors in file $1$DGA2:[ORACLE.V9.ADMIN.YUSDB.BDUMP]CAPL22_YUSDB2_BG_LMS0_006.TRC;:
ORA-00484: LMS* process terminated with error
6-AUG-2008 06:30:37.40:
System state dump is made for local instance
6-AUG-2008 06:30:37.56:
Trace dumping is performing id=[cdmp_20080806063037]
6-AUG-2008 06:30:42.47:
Instance terminated by PMON, pid = 20401500
Starting up ORACLE RDBMS Version: 9.2.0.2.0.
Oracle Bug? who can help me? Thanks in advance.
Message was edited by:
Efrain

> I use 100M switch not Gibit switch, maybe it's also problem?
Could be.. but one will need to look at the interface stats and switch stats to see whether the interconnect is coping.
For 2 nodes, a decent interconnect is not that expensive. 2 HBC cards and small Inifiband switch.
RAC is only as robust as the hardware it is running on.

Similar Messages

I changed ISPs and had to change the server settings. Now one of my two email addresses crashes every time I try to read the emails.

At first Thunderbird downloaded every email from the problem account going back to 2011. I changed the server settings to only download emails from the last 15 days and to delete any messages older than 15 days.
I can send emails from this account and my other email account works fine for sending and reading mail. However, this problem account crashes Thunderbird every time I try to read the messages. There are now only 300 messages in the inbox and I've been able to have them marked as read. But I still can't read any of them.
What's the fix on this?

Try to start *Windows* in safe mode with networking enabled.
- Win8 http://windows.microsoft.com/en-US/windows-8/windows-startup-settings-including-safe-mode
- Win7 http://windows.microsoft.com/en-US/windows7/Start-your-computer-in-safe-mode
- XP http://www.microsoft.com/resources/documentation/windows/xp/all/proddocs/en-us/boot_failsafe.mspx
Still in Windows safe mode, start Thunderbird in safe mode.
- https://support.mozilla.org/en-US/kb/safe-mode
Does the problem go away?

One instance and two databases?

Hello, I've installed oracle database on my server. I've read a lot of documentation to learn the basic concepts.
Ok, I have a question. I have read that one instance only can manage ONE database. Is it true? If it's true, how on earth I can create two database? for example, I want to create database-one and database-two.
I would create these with CREATE DATABASE command. How does Instance know if I want to access to database-one or database-two. How I indicate it in initilitation parameter file of my instance?
In sumary, Can one instance works with two databases?

Look at the following output. If I do not set my ORACLE_SID variable, the 9.2.0.7 client connects to the 9.2.0.7 database instance. If I set my ORACLE_SID , the same 9.2.0.7 client now connects to the 10.2.0.1 database (the database specified in the ORACLE_SID variable).
C:\>set ORACLE_SID=
C:\>sqlplus kkishore/kkishore
SQL*Plus: Release 9.2.0.7.0 - Production on Mon Jul 17 09:27:55 2006
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Connected to:
Oracle9i Enterprise Edition Release 9.2.0.7.0 - Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.7.0 - Production
SQL> exit
Disconnected from Oracle9i Enterprise Edition Release 9.2.0.7.0 - Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.7.0 - Production
C:\>set ORACLE_SID=ORA10GR2
C:\>sqlplus kkishore/kkishore
SQL*Plus: Release 9.2.0.7.0 - Production on Mon Jul 17 09:29:31 2006
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL> disconnect
Disconnected from Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL>Message was edited by:
Kamal Kishore

One instance for two different application

Hello,
I have to solve following problem..
need have one instance of object for two different application..fist is standard SE application and second is EE (web)..
SE application is a server..listen on and accepting connection from mobile phone..and EE application need to show list of connected devices and should say to SE application that new device was connected..
I would like to do this without database..only server and web app..
I have thought about RMI..but hope there is more simple solution..
Thanks a lot for your ideas.
Dmitrij

I would like to do this without database..only server and web app..You have your reasons for wanting this to be the case. If you are sure you don't want a DB, then pass the state of the object between the two ends using XML as the transport form. As someone said, you can use JMS or you can write an HTTP POST or probably other transport mechanisms other people will suggest to actually get it over the wire from one side to the other.
As to your question of how to sync state, just make the object you're sharing final after it's been created (on the side that creates it) and make the reference that receives it and holds it final also, after you've reconstructed it from the XML you received.

Why are there two ns-httpd processes running for one instance of iws 6

Hi all,
I'm used to iWS 4.x where there's only one watchdog and one ns-httpd process running for any given install of iWS. Now in iWS 6.x I noticed there are two processes for the admin server and two processes for the actual single web server instance, along with two watchdogs.
What are the extra web server instances doing? I noticed that there is a 'parent' relationship with the pid's, so I guess one is spawning the other, but why? If it's to have persistent state over restart, isn't that what uxwdog is for?
Thanks in advance,
James Huang

goldeye wrote:
$ ps -ef|grep ns-httpd
root 5612 5611 0 May 09 ? 2:22 ns-httpd -d /u00/iplanet/server6/https-admserv/config
root 5613 5612 0 May 09 ? 5:30 ns-httpd -d /u00/iplanet/server6/https-admserv/configThis is Administration Server
root 22940 22939 3 May 16 ? 155:11 ns-httpd -d /u00/iplanet/server6/https-eagnmnsu24f.xxxx.gov/config
nobody 23040 22940 0 May 16 ? 136:35 ns-httpd -d /u00/iplanet/server6/https-eagnmnsu24f.xxxx.gov/configThis is Web Server instance.
Can you send the output of ps -ef|grep iplanet ?
iplanet Web Server is very old please move to latest Web Servers.
You can download latest Web Servers from [http://www.sun.com/download/index.jsp?cat=Web%20%26%20Proxy%20Servers&tab=3&subcat=Web%20Servers].
Please use either of the two
Sun Java System Web Server 7.0 update 5 (release notes are located @ [http://wikis.sun.com/display/WebServerdocs/Release+Notes]) or
Sun ONE web Server 6.1 SP11 ( release notes are located @ [http://docs.sun.com/app/docs/doc/820-7659]).
Migration Guide will help you in migration [http://wikis.sun.com/display/WebServerdocs/Installation+and+Migration+Guide] or let us know if you have any problems.
When I start Sun Java System Web Server 7.0 or Sun ONE 6.1 Server instance, I can see one watchdog and two webservd processes :
$ps -ef|grep /export2
user 7967 7966 ... webservd -r /export2/ws -d /export2/ws/https-....
user 7966 7965 ... webservd -r /export2/ws -d /export2/ws/https-...
user 7965 1 ... ./webservd-wdog -r /export2/ws -d /export2/ws/https-...

Which is better? Two Oracle user in one instance OR in two instances?

Which is better?
I could not find any benchmark for number of instances per user
in the same machine.
suppose that you have two major Oracle user, from performance
point of view it is better to make separate instances in same
machine or keep these two users in the same instance.

Hi.
I understand that you will use oracle for two different
applications on the same host.
Each instance has its SGA and background processes. So, using
two different schemas(users) in one instance you can share
phisical memory between the applications. Concerning background
processes, you can start as many of them as you wish, so it's
not a problem. If you decided that one DBW is not enough, you
could configure oracle to use two or more.
best regards,
Andrew

ICal crashes when editing one instance of recurring event

Maybe it is just my imagination, but this started just as MobileMe was released.
If I edit or delete one instance of a recurring event, iCal crashes about 90% of the time. I'm not sure why it does work in a very few cases. Usually it will show as "Not responding" in my activity monitor and start using all memory. When I force quit, the log files include the following:
Command: iCal
Path: /Applications/iCal.app/Contents/MacOS/iCal
Version: 3.0.4 (1263)
Build Version: 3
Project Name: iCal
Source Version: 12630000
Parent: launchd [458]
Any tips? My iCal is nearly unusable as most of my events are recurring, and editing events is a necessary tool. Let me know if I should post more of the error log.

Hi,
the following might help:
http://discussions.apple.com/message.jspa?messageID=7761861#7761861
It fixed the problem for me.
Good luck!

One instance registered in two rman catalogs is possible?

Customer wants that one instance be registered in two different rman catalogs.
and perform two different backups.
There is a sincronism between control file and Catalog, what could mess it up having two catalogs. On the other hand there is the DBID who identify the database on each catalog.
Can anyone clarify if this is possible or not?
TX

Lalbuque,
RMAN performs a backup using a "Media Manager Layer" or "MML". There can only be one MML installed at once on a particular ORACLE_HOME and it seems that yours is TSM; you cannot run an RMAN backup to tape without an MML. This is true even if you backup WITHOUT a catalog so adding a second will not help.
There are a number of solutions to this problem. Here are a couple.
Firstly, you could backup your database to the non-TSM tapes without RMAN. You can do this either by shutting the database down first or you could put the database into backup mode for the duration of the backup. The former is simpler but a) requires an outage and b) requires you to be absolutely sure that your shutdown was clean. The latter requires some detailed knowledge of oracle backup concepts but avoids the other problems.
The second solution is to backup your database to disk with RMAN and then to copy the resultant files to tape.
Whatever solution you choose please bear three things in mind:
1) The purpose of backup is to recover. Make sure you have a documented recovery process and test it.
2) The use of a second catalog is not required.
3) Test your ability to recover. Have I said that already :-)
--Tony Hasler

Central CCMS Alert for Java instance server node crash

Hi,
Is it possible to trigger an alert of an Java system server node crash using central CCMS alerts.
J2EE instance CCMS alert does show some MTEs, however they do not trigger alerts. I got this info from below link
http://help.sap.com/saphelp_nwce711/helpdata/en/46/11aaf352da14dce10000000a155369/frameset.htm
Any idea how to trigger the alert for server node crash.
I understand that server node crash should be investigated for permanent fix, however we need this as a proactive measure to know the crash if they happen.
Thanks
Imtiaz

Hi Imtiaz,
Do you see any error on the view>status auto-reaction? Have you been able to assign an auto-reaction to this MTE?
Cheers,
Maurício

Oracle 9i database is not mounted on one instance

platform: IBM AIX
db version: oracle 9i two node rac
due to network problems one of the instance is down,after server restart the database is not mounting.just it is hanging.alert log has no information.
please help

Nagendra T wrote:
platform: IBM AIX
db version: oracle 9i two node rac
due to network problems one of the instance is down,after server restart the database is not mounting.just it is hanging.alert log has no information.What is the status of the instance, You can check from other node(instance) also.
If try to startup instance, any errors from SQL?
No information in alert log file? Are you chcking in right alert logfile?
Check any background processes are running or not?
$ps -ef|grep pmon
However one instance running, So you can kill that session which is not running and start again.

ORA-00488: RBAL process terminated with error causing ASM instance to crash

Hi Freinds,
I am manging a three node ASM cluster on 10g(10.2.0.2).Yesterday I observed one of the ASM instance got crashed .On reviewing the Alert log of crashed ASM instance I found
"*ORA-00488: RBAL process terminated with error*"
and subsequently
"*PMON: terminating instance due to error 488*".
I could restart the crashed instance without an issues but I want to know what kind of event could lead to such scenario.
I searched net and metalink but didn't found any helpful information.
Did any of you faced similar issues and do you have understanding why and when it happens.
Thanks in Advance
Arindam.

Hi,
There is bug reported on such issue, please check further on metalink if you have support.
Dilip Patel.

One machine = One Instance

Hi there,
How can i prevent that more than one instance of my application from being created in the same machine. How can i rename the process associated to the application (instead of the usual javaw.exe)?
Thanks for any help you can give me,
Carlos Ferreira

When I was writing for Windows, we did this via global
Mutex objects. Your app (on startup) would attempt to
create and get ownership of a global named Mutex
(usually the App name). If you got ownership, you were
the only one running.
The problem here with Java is that you MIGHT be
running multiple JVMs. Without access to a machine
level Mutex, it is difficult to ensure One instance
per machine (that behaves well following a crashor
error).Um - you do realize that the suggested socket approach guarantees exactly this? I come up, and either I get the socket or I don't. If I fail - someone else is already running, exit. When I leave, or if I crash, the socket is automatically freed up for me.
(And, having had to debug poorly-written Windows Mutex code, I know that mutexes can be left in odd states if the holder falls over while holding them. It's easy to code to notice and deal with this - but if you fail to do so, then your app fails to run again, unless/until you reboot the machine.)
However, one could make use of JMS to simulate this.
Lots of overhead to set up the client and server ends
of this simply to ensure 1 application per machine,
but it should work. We do have to assume that the
application plays by the rules and attempts to
validate its one-ness prior to startup.Ow. JMS is a hugely-overcomplicated tool to use for this straightforward application.
And, regarding everyone "playing fair" - well, yeah, that's a given. But since what we're talking about here is two instances of the same program, then it pretty much by definition either plays the game, or it doesn't.
Grant

How to run multiple Instances of wls5.1 under one instance

Hello forum,I want to make 3 instance of weblogic that runs under one instance,but
i don't know how.
for example I will like to have the following.
Weblogic is my main instance,so i will like to have more instances called web,db,batch
witch run on their own under.
Weblogic/web,db,batch,
I have made the 4 instance first weblogic has empty weblogic.properties file,
and the otheres 3 have their full weblogic.properties,my problem is how can I
refer eatch instance individually, that means the
subinstances like web,db,batch in my startweb.cmd,startdb.cmd,startbatch.cmd
I don't know how to start eatch instance alone.
HELP

To run more than 1 instance of Logic at the same time, you really need 2 copies as while the license allows you to install on 2 machines (1 laptop 1 desktop ideally), they can not run at the same time.
So, where now? Well there is a legitimate way around this, assuming you have a fast enough network between the 2 computers and that is to have 1 machine running as a Logic Node to share the work out. When I'm not 'Out and about' I could use the MacBook Pro as a node for my home setup (For example) running some of the soft synths for example.

Can i create more than one instance for single database?

HI
anybody can tell me how to create more than one instance for single database(without using RAC)?
thanks
Kalanidhi

Anandji,
I sometimes think differently. I also think like what this person is asking..(although I have installed and administering 5 different RAC databases)
The main doubt of mine is that we create multiple instances accessing the same storage area.
Anyway the controlfiles of all instances will have the same information except the instance_names and instance id's.
Hence I always think , why can't it be possible to to create another instance in the same system and point it to a controlfile which is copied from the other instance.
Why can't the same files be shared between the two instances in a same system ?
I know I am thinking a little bit confusing...
Anyway I need to test that also...
Mahesh.

How to install more than one instances on one machine

Hi ALL,
I want to install more than one instances on one machine, but i am a freshman on SAP field. How and where can i get the related infomation for that? thank you very much !!

Hi tieyan,
Glad the information was of some help.
For Windows, there is an older note <a href="http://service.sap.com/sap/support/notes/28392">28392</a> that has some more information. It is important also that the SAP parameter PHYS_MEMSIZE is sized correctly according to note <a href="http://service.sap.com/sap/support/notes/88416">88416</a> (the bottom of the note gives some examples). For example, if there were two central instances and the two Oracle database instances on one host, PHYS_MEMSIZE for both instance profiles (of the central instances) would be 35% of the RAM (1/2 of 70%). In the case of MSSQLSRV 2000, for PHYS_MEMSIZE you would refer to the scenarios of note <a href="http://service.sap.com/sap/support/notes/327494">327494</a>.
If you ever want to thank somebody else for helping in answering a question, the <a href="https://wiki.sdn.sap.com/wiki/display/HOME/RulesofEngagement">Rules of Engagement</a> has a section 'Thank Others by Giving Points' that explains this further.
Regards, Mark

One Instance of two instance/nodes Crash in RAC

Similar Messages

Maybe you are looking for