Problem to cluster novell-named dns

I have the novell-dns running on one node and I try to make it cluster enabled on an NSS-pool.
I follow the documentation which is very confusing.
It's really not clear, if we have to create a LUM-enabled user called "named" with UID 44 or just keep on using the local "named" user ?
Actually if you create a lum user with the same name and uid as a local user it gives problem...
And if you use the default user, it does not work, the rights are not set up on the NSS pool.
Does anyone already clustered a novell-named DNS ?
Thanks
Sylvain

Hello,
Don't be sorry you were very helpful.
I think I have finally found a solution to make it work:
- Create the DNSVOL on NSS shared pool
- Set the Novell DNS on every nodes on the cluster ( do not create DNS servers )
- Create a LUM user "novellnamed" for example ( use default UID, primary group DNSDHCP-GROUP)
- run the novell script ncs_sh with correct parameters
- Make sure the LUM user has file trustee rights on /media/nss/DNSVOL/etc/opt/novell/named
- Create a DNS server for the NCP virtual server of the NSS shared pool.
- Then add the load and unload command in the cluster resource script, be careful to mention the novell volume name after the -V option and not the nss path: /opt/novell/named/bin/novell-named -u novellnamed -d 5 -V DNSVOL
- If you have errors in the /var/opt/novell/log/named/named.run, check that the LUM user is owner with sufficient rights in :
/var/opt/novell/run/named & /var/opt/novell/log/named.
This worked fine for me, but the documentation is definitely very confusing.
Thanks and very happy new year !
Sylvain

Similar Messages

  • Novell-named failed to start / 2nd DSFW server

    Posting this here in case anyone else runs into this and it took me some tinkering to figure it out.
    I was busy installing a 2nd DSFW server and ran into a named error during the provisioning where it would start successfully but fail in the 'is it really running' check.
    When manually starting novell-named it gives a successful start but then exits after a second or so.
    When checking /etc/opt/novell/named/ it turned out that there was no named.conf even after configuring the dns server from imanager.
    /var/opt/novell/log/named/named.run had
    Code:
    07-Mar-2012 16:26:07.277 general: main: notice: starting BIND 9.3.2 -u named
    07-Mar-2012 16:26:07.319 general: dns/db: critical: Unable to login Error code:-223
    07-Mar-2012 16:26:07.320 general: dns/db: critical: Failed to load RRs of rootserver zone with error -112
    07-Mar-2012 16:26:07.320 general: dns/hints: warning: Loading Root data from directory Failed
    07-Mar-2012 16:26:07.321 general: server: info: loading configuration from '/etc/opt/novell/named/named.conf'
    07-Mar-2012 16:26:07.321 config: isccfg/parser: error: none:0: open: /etc/opt/novell/named/named.conf: file not found
    07-Mar-2012 16:26:08.359 general: dns/db: critical: Unable to login Error code:-223
    07-Mar-2012 16:26:08.359 network: interfacemgr: info: dns_edir_get_multival has returned error inside store_dnsserver_ip_address:25
    07-Mar-2012 16:26:08.359 network: interfacemgr: error: Error occured while updating the IP list of the DNS server object:25
    07-Mar-2012 16:26:08.359 general: server: critical: loading configuration: file not found
    07-Mar-2012 16:26:08.359 general: server: critical: exiting (due to fatal error)
    What I ended up doing was copied the options { } bit from the already up and running DSFW server into the new one, adjusted the edir object names and then novell-named would start and keep running.
    It was however still giving login errors in the logfile, after remembering it used the servers commonproxy user I ran '/opt/novell/proxymgmt/bin/change_proxy_pwd.sh -A yes' (you can find this in crontab already) and then everything started to work for real after restarting novell-named.
    Somewhere along the line it lost the correct password for the proxy user but there is hardly any indication of this, it gave me enough searching around to post this :-)
    The manual creation of the named.conf file is *probably* not needed, but mentioning it just in case.
    also, it is safe to ignore the 'unable to authenticate to ldap with <insert AD with DC= credentials here>' error you get when your eDir ldap server you are using to install eDir is not a DSFW server, this gave me a 'hmmmm' moment too.

    It most likely is a problem with the casa credential. It could be the credentials are there but their is a mismatch with with the Common Proxy user's password.
    You can verify the common proxy user's password with
    common-proxy-casa-repair-tool
    Then run the novell-dns-casa-repair-tool to sync up the dns casa keys with the correct credentials.

  • Novell-named Unable to lock file /etc/nam.conf..?!?

    Getting this error on startup of novell-named
    >>
    # rcnovell-named restart
    Shutting down name server BIND waiting for novell-named to shu(28s) done
    Starting name server BIND Unable to lock file /etc/nam.conf.
    Permission denied
    <<
    Seems superficial, as dns seems to work fine. However, wondering what/why/etc. Can't think of a relationship...
    Anyone any suggestions?
    Cheers
    David

    djbrightman,
    It appears that in the past few days you have not received a response to your
    posting. That concerns us, and has triggered this automated reply.
    Has your problem been resolved? If not, you might try one of the following options:
    - Visit http://support.novell.com and search the knowledgebase and/or check all
    the other self support options and support programs available.
    - You could also try posting your message again. Make sure it is posted in the
    correct newsgroup. (http://forums.novell.com)
    Be sure to read the forum FAQ about what to expect in the way of responses:
    http://forums.novell.com/faq.php
    If this is a reply to a duplicate posting, please ignore and accept our apologies
    and rest assured we will issue a stern reprimand to our posting bot.
    Good luck!
    Your Novell Product Support Forums Team
    http://forums.novell.com/

  • Novell OES DNS not forwarding

    Never mind: it works, I just needed to wait for the dynamic config. Please delete.

    On 04.12.2012 05:26, mpatterson2100 wrote:
    >
    > I am having a problem getting novell dns to foward to another dns
    > server. Even though I have an entry in the forwarding list, my server is
    > not forwarding queries to that server.
    >
    > Here is a summary of my configuration:
    > Server01:
    > Services: (Novell DNS, Novell eDirectory)
    > IP address: 192.168.100.1
    > zones: test.local, 100.168.192.IN-ADDR.ARPA
    > forwarding list: 10.2.1.141
    > OS: OES 11 SP2, SLES 11 SP1
    Basic stuff first, as this ofte nis a problem: YOu *are* running
    novell-named, *not* named, yes?
    The usual cause for forwarding not happening is missing/corrupt
    rootserverinfo zone. Please check or post your dns server logs.
    CU,
    Massimo Rosen
    Novell Knowledge Partner
    No emails please!
    http://www.cfc-it.de

  • SAPLPD problem on cluster

    Hello experts,
    On our BI 7.0 system, we have our Dev and QA environments on single servers. However, for the Production system we use Microsoft Clustering (MSCS).
    I am not sure if my problem is related to the cluster architecture at all, but I am having problems printing to printer named LP01 which is supposed to call the SAPLPD program and pass the print job to the default printer as defined in my Windows XP desktop OS.
    The LP01 printing works perfectly when I am printing to LP01 from my Dev and QA machines, but on the Prod server if I use LP01 it "doesn't call" the SAPLPD program -- I don't see the SAPLPD listener popping up on my screen and transferring the job to the OS.
    Can you please help me with this?
    Thank you in advance.

    hi ugur,
    what kind of cluster arcithecture you're using ? Is your DB and CI reside on same server or different server ?
    If it is reside on same server, make sure that this server has minimumly one spool WP. If you're CI and DB is reside on different server, just make sure both server have minimumly one spool WP.
    ardhian
    http://ardhian.kioslinux.com
    http://sapbasis.wordpress.com

  • How to name elements and create cluster of named elements

    I wanted to detail the steps I found, so that others could more easily follow what I found going through these discussions.  While there are numerous examples, they lack details such as function names and how to achieve the steps.  Most of them seem to cover only arrays.  I welcome corrections.  I'm not saying this is the best way; it is, however, a simple detailed description of one way to do it.
    I wanted to create a cluster of various data types that could then be manipulated by name.  Unfortunately, simply wiring an existing cluster to a Bundle by Name.vi does not assign the existing labels of the elements to the name field of the cluster.  I don't know why.
    1.  Create an input cluster with dummy constants of the proper data type for each element.
    How do you create an input cluster?  Searching functions finds no "Input Cluster" function.
            The only way I found was to place a dummy Bundle by Name.vi , right-click at the top of the leftmost of the two rectangles and select Create, then Constant. The result looks like the attached Creating Input Cluster.PNG
             You might as well delete the Bundle by Name.vi at this point, because the changes make to the input cluster will only be reflected when it is connected to a NEW Bundle by Name.vi
             Next, for my purposes, I needed constants for DBL numeric, string, and array of strings.  I started by placing  one DBL and one string constant on the block diagram.  I copied these to create the ones I needed. Make the labels visible and type in the names corresponding the Bundle Names desired.  
             Since there is no function called "String Array constant", it is necessary to place the "Array Constant" function on the Block Diagram, then drag a string constant into the empty right hand square of the Array Constant.  It immediately changes color to show it is a string array.
              Drag each of these constants into the input cluster so that they appear in the order desired.  It is a bit of a pain because the square resizes tightly around the existing elements after each insertion.  You have to keep expanding the box at the bottom by dragging its sizing handles, so you can be sure each item is added to the bottom of the list.  When done, my input cluster looks like Completed Input Cluster.png attached.
    2.  Place a NEW Bundle by Name.vi on the Block Diagram.  Wire the input cluster to the top of the left rectangle of the function. Drag the down sizing handle of the Bundle by Name.vi until all the elements of the cluster are visible.  You will see the names and it will look like Bundle by Name with assigned names.png attached.
    3.  All that remains is to connect the data wires for the individual elements to the left side of each named block and wire the output cluster.  The result is shown in Complete design producing cluster of named elements.PNG attached. (Since only three attachments seem to be allowed, this last will be in a follow-up post.)
    Attachments:
    Creating Input Cluster.PNG ‏3 KB
    Completed Input Cluster.PNG ‏3 KB
    Bundle by Name with assigned names.PNG ‏8 KB

    I wanted to add a brief statement about EDITING cluster controls because newcomers to cluster usage (like myself).  The HELP is good on creation but says nothing about modification.  While it has been posted other places, I think a "newbie" might appreciate getting this info here.
    To edit a cluster control, open the CTL file, then
    Add a control by creating the control on the front panel, then save the control.
    To remove a a control, delete the control, then save the control file.
    (Simple enough)
    To modify an existing control
    1.  Move the control out of the cluster into a blank VI.  
    2.  SAVE the cluster control file (CTL).  
    3.  Modify the control in the blank VI.
    4.  Drag the modified control back into the cluster control.
    5.  SAVE the cluster control file (CTL)
    NOTE WELL: don't skip step 2. 

  • Novell-named on DSfW servers leaks memory

    Hi,
    before I open a SR, does this ring a bell for anyone? Any configuration
    pitfalls?
    Here the symptoms are that on all our 3 Xen-VMs running DSfW (OES11SP1
    fully patched) novell-named reaches ~2GB virtual memory usage after 14
    days and usually dies a few days later because of OOM (the VMs have 4GB
    assigned). Only workaround I found so far is to set the cache size to 0,
    but that's not what we want.
    Any ideas?
    Franz.

    Originally Posted by Franz Sirl
    Hi,
    before I open a SR, does this ring a bell for anyone? Any configuration
    pitfalls?
    Here the symptoms are that on all our 3 Xen-VMs running DSfW (OES11SP1
    fully patched) novell-named reaches ~2GB virtual memory usage after 14
    days and usually dies a few days later because of OOM (the VMs have 4GB
    assigned). Only workaround I found so far is to set the cache size to 0,
    but that's not what we want.
    Any ideas?
    Franz.
    Hi Franz,
    We are able to reproduce this issue in our local scale environment and have fixed the same in the forthcoming OES11 SP2 release. The same will be back ported to OES11 SP1 patches. You can expect the fix in the next OES11 SP1 patches.
    Thanks,
    Praveen Kumar

  • Problem associated with Novell Filesystem?

    Hi,
    We are currently upgrading a financial software in our organization and this software is running from a Novell share for the ressources files. Since this upgrade it takes about 3-4 mins before the program starts. The representative from this compagny says that it is due to Novell Filesystem...
    We validated this theory by doing the share on a Windows server (SAMBA) and we had no problem with it.
    The current patch is that we have to copy localy two ressources files on every computers that use this software.
    Theses files are .obj files that are actually zip files with no compression that contains each about 85000 files +
    We tried a few settings with the Novell Client... like file caching and other caching settings... but had no luck with it...
    So is there really a problem with the novell filesystem ?
    Can we change any settings to fix this in remote manager or imanager ?
    Thank you

    On 24/02/2012 21:16, anto28 wrote:
    > We are currently upgrading a financial software in our organization and
    > this software is running from a Novell share for the ressources files.
    > Since this upgrade it takes about 3-4 mins before the program starts.
    > The representative from this compagny says that it is due to Novell
    > Filesystem...
    >
    > We validated this theory by doing the share on a Windows server (SAMBA)
    > and we had no problem with it.
    >
    > The current patch is that we have to copy localy two ressources files
    > on every computers that use this software.
    > Theses files are .obj files that are actually zip files with no
    > compression that contains each about 85000 files +
    >
    > We tried a few settings with the Novell Client... like file caching and
    > other caching settings... but had no luck with it...
    >
    > So is there really a problem with the novell filesystem ?
    >
    > Can we change any settings to fix this in remote manager or imanager ?
    This forum is for queries relating to Novell's Dynamic File Services
    product and doesn't cover the above.
    Since you haven't said whether you're using NetWare or Open Enterprise
    Server (Linux), one of the Storage-related sub-forums @
    http://forums.novell.com/novell/nove...rprise-server/
    would probably be the best place for you to repost this.
    HTH.
    Simon
    Novell/SUSE/NetIQ Knowledge Partner
    Do you work with Novell technologies at a university, college or school?
    If so, your campus could benefit from joining the Novell Technology
    Transfer Partner (TTP) program. See novell.com/ttp for more details.

  • Easier way to convert array to cluster with named variables?

    I have an array of variables, I would like to essentially unbundle that array to named variables.
    I believe the only way of doing this is by converting the array to cluster (besides removing each element one by one)?
    Is what I have done below the best way of doing it? seems a bit weird having to unbundle and rebundle by name.
    comments appreciated.
     

    Your problem code is actually right before the red square. Instead of the "array-to-cluster-unbudle dance", you should simply use a plain index array resized for the same number of outputs. same functionality, less convoluted.
    Your wiring is a bit of a mess and it is hard to tell if the order of elements in the array is the same as in the cluster.  Maybe a simple typecast to the cluster would work.
     

  • Serious bug: call set-up problem in case of several DNS SRV records

    Hello Cisco,
    We have a MCU that consists of two servers in cluster. We have SIP SRV DNS records that point to both servers with equal priority and weight.
    All applications work nice with such setup, except from Free Jabber. Jabber is unable to set up the connection most of the time. One time the connection is successful and 5, maybe even 10 times it is unsuccessful.
    For testing, we removed SIP DNS records pointing to one server. This way Jabber works much better. There are some cases when the call set up fails but in most cases it works.
    Looking the logs of the MCU, we can see three different ways, how call set-up may fail. It is probably unreasonable describe the details in this forum message. Anyway, it seems to be sure that in case there SIP SRV records point to one server then Jabber is able to connect the MCU, in case the records point to two servers equally then Jabber is pricnipally unable to connect the MCU. This bug should be fixed, IMHO.
    Btw, what record does Jabber follow, is it _sips._tcp or _sip._tls?
    Greetings and thank you in advance,
    Marko Laurits

    Hello Cisco,
    We have a MCU that consists of two servers in cluster. We have SIP SRV DNS records that point to both servers with equal priority and weight.
    All applications work nice with such setup, except from Free Jabber. Jabber is unable to set up the connection most of the time. One time the connection is successful and 5, maybe even 10 times it is unsuccessful.
    For testing, we removed SIP DNS records pointing to one server. This way Jabber works much better. There are some cases when the call set up fails but in most cases it works.
    Looking the logs of the MCU, we can see three different ways, how call set-up may fail. It is probably unreasonable describe the details in this forum message. Anyway, it seems to be sure that in case there SIP SRV records point to one server then Jabber is able to connect the MCU, in case the records point to two servers equally then Jabber is pricnipally unable to connect the MCU. This bug should be fixed, IMHO.
    Btw, what record does Jabber follow, is it _sips._tcp or _sip._tls?
    Greetings and thank you in advance,
    Marko Laurits

  • Can we install a new mssql cluster on the same windows cluster which already containes a mssql cluster with named instance

    We have a MSSQL 2008R2 Enterprise edition with a two node active passive fail-over cluster running on 2008R2 windows cluster with out any issues,
    Now my question is can we add one more MSSSQL cluster instance for the same setup with out disturbing the existing one ?
    Also give thoughts on load sharing as the second node is mostly ideal now except fail-over scenarios,
    Why we go for this situation is because of the collation setting which can be set only one per instance(Database collation setting change not working), we need a different default collation for the new setup

    hi,
    >>Now my question is can we add one more MSSSQL cluster instance for the same setup with out disturbing the existing one ?
    Yes it is possible .You need to add new drives as cluster aware and install SQL server and put data and log files on thse drives.YOu would need to create named instance of SQL server and need to create different resource group.Both old installation and new
    onw would work separately.
    >>Also give thoughts on load sharing as the second node is mostly ideal now except fail-over scenarios,
    Good point indeed.You are about to create Multi instance cluster and should plan for scenario where one node is down and other node is handling load for both instances.Memory and CPU should be enough to handle the load.
    >>Why we go for this situation is because of the collation setting which can be set only one per instance(Database collation setting change not working), we need a different default collation for the new setup .
    Just for collation if you are installing new instance seems little wierd to me.You can manage collation at column ,database and at server level.
    http://technet.microsoft.com/en-us/library/aa174903(v=sql.80).aspx
    Please mark this reply as the answer or vote as helpful, as appropriate, to make it useful for other readers

  • Problem: Stopping cluster due to unhandled exception .. Unable to refresh

    Greetings
    While testing a rolling upgrade of an application that uses Coherence 3.5.2 on 62-jvm cluster hosted on
    11 physical machines, we encountered a situation where after the upgrade was completed, most jvms
    in the system abruptly left the cluster. The physical hosts are running CentOS 5.4, the java used is
    the 64-bit server version 1.6.0_16-b01. The test was run under a scenario that imposed "moderate"
    load, with cpu usage on the physical machines never exceeding 60% busy, network bandwidth never
    exceeding 5%, and with some free physical memory. Swapping did not occur during any time during
    the test. I believe we are using the default tangosol-coherence.xml. We got the error below in
    our coherence.log files on all of the systems, all at about the same time (within 10 milliseconds).
    55 of the jvms left the cluster during the incident. There were 55 copies of the error message in
    the various logs, all nearly identical except for the time and member id.
    My questions include
    - what does the error mean?
    - what could cause it? (I investigated system logs, and found no evidence of the NIC cards
    going off line at the time. Any suggestions about how to look for evidence of broadcast storm?)
    - how can we keep it from happening again?
    Many thanks for your help -
    Mike Murphy
    2011-04-26 17:34:14,629 Coherence Logger@9224544 3.5.2/463 ERROR 2011-04-26 17:34:14.629/1929.311 Oracle Coherence GE
    3.5.2/463 <Error> (thread=PacketListenerN, member=30): Stopping cluster due to unhandled exception: com.tangosol.net.mes
    saging.ConnectionException: Unable to refresh sockets: [UnicastUdpSocket{State=STATE_OPEN, address:port=10.48.88.116:809
    1}, MulticastUdpSocket{State=STATE_OPEN, address:port=224.3.5.2:10013, InterfaceAddress=10.48.88.116, TimeToLive=4}, Tcp
    SocketAccepter{State=STATE_OPEN, ServerSocket=10.48.88.116:8091}]; last failed socket: MulticastUdpSocket{State=STATE_OP
    EN, address:port=224.3.5.2:10013, InterfaceAddress=10.48.88.116, TimeToLive=4}
    at com.tangosol.coherence.component.net.Cluster$SocketManager.refreshSockets(Cluster.CDB:91)
    at com.tangosol.coherence.component.net.Cluster$SocketManager$MulticastUdpSocket.onInterruptedIOException(Cluste
    r.CDB:9)
    at com.tangosol.coherence.component.net.socket.UdpSocket.receive(UdpSocket.CDB:33)
    at com.tangosol.coherence.component.net.UdpPacket.receive(UdpPacket.CDB:4)
    at com.tangosol.coherence.component.util.daemon.queueProcessor.packetProcessor.PacketListener.onNotify(PacketLis
    tener.CDB:19)
    at com.tangosol.coherence.component.util.Daemon.run(Daemon.CDB:42)
    at java.lang.Thread.run(Thread.java:619)
    Caused by: java.net.SocketTimeoutException: Receive timed out
    at java.net.PlainDatagramSocketImpl.receive0(Native Method)
    at java.net.PlainDatagramSocketImpl.receive(PlainDatagramSocketImpl.java:136)
    at java.net.DatagramSocket.receive(DatagramSocket.java:712)
    at com.tangosol.coherence.component.net.socket.UdpSocket.receive(UdpSocket.CDB:20)
    at com.tangosol.coherence.component.net.UdpPacket.receive(UdpPacket.CDB:4)
    at com.tangosol.coherence.component.util.daemon.queueProcessor.packetProcessor.PacketListener.onNotify(PacketLis
    tener.CDB:19)
    at com.tangosol.coherence.component.util.Daemon.run(Daemon.CDB:42)

    Hi Mike,
    It looks like you are having problems with multicast. You can run the multicast test described here
    [http://download.oracle.com/docs/cd/E15357_01/coh.360/e15723/tune_multigramtest.htm]
    which will help in diagnosing the problem

  • Problems with Cluster Ready Service installation (RAC) in Oracle 10g R1

    Hello
    I have got problems with this installation. I am starting Oracle user
    At the end of the installation I get a mesage when all Configuration Assistants are failed
    I have got 2 computers with Windows 2000 PRO, two networks (private and public)
    As a shared disk I use one external disk (with USB) attached to one computer.
    What to do here?
    please for a reply
    Martin
    PROT-1: Failed to initialize ocrconfig
    SCLS error 1 reading current privileges.
    Internal Error Information:
    Category: 1234
    Operation: scls_iddb_has_privgrp_by_name
    Location: NetLocalGrou
    Other: get local group failed
    Dep: 0
    Step 1: checking status of CRS cluster
    Step 2: configuring OCR repository
    ignoring upgrade failure of ocr(-1)
    failed to configure Oracle Cluster Registry with CLSCFG, ret 102
    Result code for launching of configuration assistant is: 1
    The OUICA command is launched from D:\oracle\product\10.1.0\crs\oui\bin\ouica.bat.
    Launched configuration assistant 'Oracle Notification Server Configuration Assistant'
    Tool type is: Optional.
    The command being spawned is: 'D:\oracle\product\10.1.0\crs/bin/racgons.exe add_config migi:4948'
    Start output from spawned process:
    End output from spawned process.
    Configuration assistant "Oracle Notification Server Configuration Assistant" failed
    Result code for launching of configuration assistant is: 1
    The OUICA command is launched from D:\oracle\product\10.1.0\crs\oui\bin\ouica.bat.
    Launched configuration assistant 'Oracle Private Interconnect Configuration Assistant'
    Tool type is: Optional.
    The command being spawned is: 'D:\oracle\product\10.1.0\crs/bin/oifcfg.exe setif -global "public"/10.0.0.0:cluster_interconnect "public"/192.168.1.0:public'
    Start output from spawned process:
    PRIF-12: failed to initialize cluster support services
    End output from spawned process.
    Configuration assistant "Oracle Private Interconnect Configuration Assistant" failed
    PRIF-12: failed to initialize cluster support services
    Result code for launching of configuration assistant is: 1
    The OUICA command is launched from D:\oracle\product\10.1.0\crs\oui\bin\ouica.bat.
    Error:*** Alert: Some of the configuration assistants failed. However these are optional assistants, so they are not required for the correct configuration of your system. If you want to try to run those assistants again, select the failed assistants and click the 'Retry' button. ***
    User Selected: Yes/OK
    Starting to execute configuration assistants
    Launched configuration assistant 'Oracle Cluster Ready Services Configuration Assistant'
    Tool type is: Optional.
    The command being spawned is: 'D:\oracle\product\10.1.0\crs/install/crssetup.config.bat'

    Also you clusterware installation installs to an ORACLE_HOME.
    Oracle does only make a differentiation, if it has to be clear, that you got a clusterware home and a database home.
    Normally if a patch is referring to $ORACLE_HOME (and the patch can be used for clusterware & database), it just means the installation directory of the oracle software installed.
    Sebastian

  • Potential dynamic server list updating problem in cluster

    We are running a cluster of four Windows 2000 Server boxes under WLS
              6.1SP2. We have seen the following behavior in our cluster. We shutdown
              a cluster member and then bring it back up. When it comes back up, the
              IIS plug-in seems to start routing all requests to that new server.
              Out cluster seems to have some multicast connectivity problems, e.g. we
              often cannot see managed servers on the admin. console, etc. What I'm
              wondering is whether the server when it comes up could be telling IIS
              that it is the only cluster member because it isn't aware of any other
              members. This would explain the plug-ins sudden affinity for the new server.
              Is this possible? Has anyone seen this?
              Thanks in advance,
              Coty
              P.S. The box that is being restarted and then hogging the requests is
              also the server where the admin. console is running. Could that be the
              issue? What exactly are the implications of running a managed server and
              the admin. console on one box?
              

    Is it that the admin server is also part of the cluster and is the server
              that is being restarted?
              Try to have the admin server for purely administrative tasks and not as part
              of a working cluster.
              If you have Debug set to ON in the plugin end, try checking the server list
              http://IISServer:port/dummyQuery?__WebLogicBridgeConfig
              This should return the dynamic and static server list that the IIS proxy is
              supposed to forward.
              Check before restart and after restart.
              Also contact support to get the newest ISAPI plugin that contains latest
              fixes.
              -Sabha
              "Coty Rosenblath" <[email protected]> wrote in message
              news:[email protected]...
              > We are running a cluster of four Windows 2000 Server boxes under WLS
              > 6.1SP2. We have seen the following behavior in our cluster. We shutdown
              > a cluster member and then bring it back up. When it comes back up, the
              > IIS plug-in seems to start routing all requests to that new server.
              >
              > Out cluster seems to have some multicast connectivity problems, e.g. we
              > often cannot see managed servers on the admin. console, etc. What I'm
              > wondering is whether the server when it comes up could be telling IIS
              > that it is the only cluster member because it isn't aware of any other
              > members. This would explain the plug-ins sudden affinity for the new
              server.
              >
              > Is this possible? Has anyone seen this?
              >
              > Thanks in advance,
              > Coty
              >
              > P.S. The box that is being restarted and then hogging the requests is
              > also the server where the admin. console is running. Could that be the
              > issue? What exactly are the implications of running a managed server and
              > the admin. console on one box?
              >
              

  • AWR report problem after cluster node switch.

    Hello all. I have some strange problem, can any one advice what to do....
    I have OracleDB (Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 on Solaris x86_64), we have two server nodes and one shared storage attached to them, db is running on node1 and if it dies db will be switched to node2, classic cluster.
    Some time ago we tested this switching, so i shut downed db and switch it to node2, and startup it there (oracle_homes are identical), every thing was ok, soo i switched it back to node1. But now i can't run awrrpt.sql or awrinfo.sql, it gives error like this:
    Using the report name awrinfo.txt
    No errors.
    create or replace package body AWRINFO_UTIL as
    ERROR at line 1:
    ORA-03113: end-of-file on communication channel
    No errors.
    ERROR:
    ORA-03114: not connected to ORACLE
    And in alert log:
    ORA-07445: exception encountered: core dump [SIGSEGV] [Address not mapped to object] [509] [] [] []
    I tried to drop AWR with catnoawr.sql and recreate it with catawrtb.sql, everything seems to be fine, but still can't run awrrpt.sql or awrinfo.sql, same error.
    Any one familiar with such problem ?
    Thanks for advice.

    I understand that I provided less than satisfactory amount of info.
    So here is more.
    I am installing the two node cluster and during scinstall one of the nodes is being rebooted
    and goes through (what I am suppose to be) an initial configuration. At the very end of the
    boot process there is a message
    obtaining access to all attached disksAt this point the boot disk activity LED is lit constantly. After some longish timeout
    the following message is printed to console
    NOTICE: /pci@0,0/pci1014,2dd@1f,2: port 0: device reset
    WARNING: /pci@0,0/pci1014,2dd@1f,2/disk@0,0 (sd1):
         Error for Command: read(10)          Error Level: Retryable
         Requested Block: 135323318          Error Block: 135323318
         Vendor: ATA                    Serial Number:
         Sense Key: No Additional Sense
         ASC: 0x0 (no additional sense info), ASCQ: 0x0, FRU: 0x0and the disk activity LED is turned off. After that nothing more happens. The system isn't
    hard hang, since the keyboard is working, and it responds to ping, but other than that
    nothing seems to be functioning.
    I understand that diagnosing such a problem isn't easy, but I am willing to invest some
    time into getting it working. I would rally appreciate some help with this issue.
    Regards,
    Cyril

Maybe you are looking for