Defining logicals on VMS

Actually we put a mechanism in just to do this very thing. On unix when
you set an env variable it is only seen by the children and we wanted to
get this exact behavior on vms. So we created a new logical name table
Forte_prctable. This is a process level logical name table but when we
issue the run command we create a com file that contains a define statement
for every logical in this table. Then when the sub/detached process executes
it first runs the command file defining local logicals as inherited by the
parent. Note that asyncronous commands actually run a detached process and so
the job table will not work for these.
To unsubscribe, email '[email protected]' with
'unsubscribe forte-users' as the body of the message.
Searchable thread archive <URL:http://pinehurst.sageit.com/listarchive/>

Hi Ella,
According to your description, you want to define security on a dimension, so that different users can see different dimension attributes or hierarchies, right?
As you said, there is no "read" and 'read/write" option on the dimension tab, we cannot set "deny" option for the dimension when creating a role. Based on my research, it seems that there is no a directly way to achieve this requirement. What we can do is
that "Deselect all members" for each attribute of the dimension in Dimension Data tab, which will hide all members of the dimension for the role. Even the dimension metadata is visible in the cube in clients, its members can't be seen. Here are some useful
links for your reference.
http://saysmymind.wordpress.com/2013/01/15/hide-a-specific-attribute-hierarchy-in-role-playing-dimension-ssas-multidimensional/
http://social.msdn.microsoft.com/Forums/sqlserver/en-US/e2596eec-7c2d-48c0-8e81-538a8c632360/hide-dimension-based-on-roles
If you have any concern about this behavior, you can submit a feedback at
http://connect.microsoft.com/SQLServer/Feedback and hope it is resolved in the next release of service pack or product. Your feedback enables Microsoft to make software and services the best that they can be, Microsoft might consider to add this feature
in the following release after official confirmation.
Thank you for your understanding.
Regards,
Charlie Liao
If you have any feedback on our support,
please click here.
Charlie Liao
TechNet Community Support

Similar Messages

ORA_OLB and ORA_UTIL logicals on VMS

As part of a conversion from an Ingress database to Oracle and the C code that
supports it, my developers asked me the following:
What VMS command procedure (.COM file) defines the logicals ORA_OLB and ORA_UTIL- both used in the
LNOCIC.COM procedure used to link OCI programs ?
Appreciate all help.
I have also posted this on the OCI forum, but I think it's more of a OS level type
question.

Funny thing: this happens in SQL Developer, but runs fine in Toad (only problem with Toad: there's no OWA Pane to see the output).Try
SQL> declare
nm owa.vc_arr;
begin
nm (1) := 1;
owa.init_cgi_env (nm);
htp.adddefaulthtmlhdr (false);
owa_util.mime_header ('application/xml', false);
owa_util.http_header_close;
-- Output the clob to the HTTP flux
htp.p ('<?xml version="1.0" encoding="Windows-1252" ?>');
htp.showpage ();
end;
Content-type: application/xml
<?xml version="1.0" encoding="Windows-1252" ?>
PL/SQL procedure successfully completed.

ORA_OLB and ORA_UTIL for OCI on VMS

As part of a conversion from an Ingress database to Oracle and the C code that
supports it, my developers asked me the following:
What VMS command procedure (.COM file) defines the logicals ORA_OLB and ORA_UTIL- both used in the
LNOCIC.COM procedure used to link OCI programs ?
Appreciate all help.

864737 wrote:
Hi John,
Maybe it's a complete nonsense but even if has no sense proceeding that way solved the issue here.
I have a query that suddenly failed to execute with the error described, version 10.2.0.3 but it does not happens in 10.2.0.4. As a workaround we did it using subselects and joining as described in the post and the query worked again.
So ORA-01445 is triggered even if it's not the cause of the error. You have there a circumstance where it fits...
So, it is clearly a bug introduced (possibly in 10.2.0.3), and fixed in 10.2.0.4.
I cannot post the query because license restrictions but believe it or not this happens with Oracle. It's great to know that the number fields of the tables does not matter. I was suprised to hear it. But your clarification was good. :D Tnx.
It seems that our DBA found that this behaviour can be corrected with an official patch.
Which clearly indicates a bug, just be sure that the patch is truly the right one.
Thank you again for your response.There are a number of bugs in MOS around ansi joins getting this error, bug
4369235.8 ORA-1445 can occur for ANSI join queries with large views/subqueries
seems possiblefor your case. However the workaround for that was to not use ansi style joins, so again, the number of columns is not directly the cause of the error.
John

Error in creating windows 8.1 VMs from templates

hello experts,
I have created a template on Windows 8.1 on SCVMM 2012 R2 (OS:Windows Server 2012) ,Now am try to create a Windows 8.1 VM on another Hyper-v Host where Windows Server 2012 r2 is installed but creation has got failed with error in job list :
Error (2915)
The Windows Remote Management (WS-Management) service cannot process the request. The object was not found on the server (hyper-v host name .com).
WinRM: URL: [http://hyper-v host name.com:5985], Verb: [GET], Resource: [http://schemas.microsoft.com/wbem/wsman/1/wmi/root/virtualization/v2/Scvmm_EthernetSwitchPortInternalSettingData?InstanceID=Microsoft:Definition\1f59a509-a6ba-4aba-8504-b29d542d44bb\Default]
Unknown error (0x80338000)
Recommended Action
Ensure that the VMM agent is installed and running. If the error persists, restart the virtualization server (hypr-v host name.com) and then try the operation again.
This error can also happen due to an older version of the VMM agent on the virtualization server. Ensure that the VMM agent is upgraded to the latest version, and then try the operation again.
please help me to resolved the issues and use created windows 8.1 templates for deploy VMs on Windows server 2012 r2 Hyper-V host.
Regards
Richa KM

I am running with an idea here that the key is here: Scvmm_EthernetSwitchPortInternalSettingData
Your virtual port profile. You are defining a setting that does not exist. (I am guessing at this since the only thing is an error message).
Brian Ehlert
http://ITProctology.blogspot.com
Learn. Apply. Repeat.
Disclaimer: Attempting change is of your own free will.

Application Administrators can't see VLAN configuration on VMs in VMM 2012 R2

Hi, I have the following issue on my VMM 2012 R2 infrastructure.
I have created a cloud with a Hyper-V Cluster in it. Configured a Application Administrator Role (Self Service) and assigned all permissions related to this cloud and Hyper-V Cluster, including a couple of VM Templates and a VM Network.
They users can see the VMs, create VMs but they cannot configure the VLAN inside the VM.
If I go to the VM settings using the full VMM admin, I can do it, select the VM Network and then select the appropiate VLAN, but when I try to do this using the Application Administrator account I configured specifically to this Cloud and Cluster, I can
select the same VM Network but can't see any options related to VLAN configuration.
What am I missing? Is this the default behavior? Do I need an additional permission? Where?
Thank you for your responses.
Regards.
Eduardo Rojas

That is correct.
Application Admins 'consume' resources (compute and storage) and have no insight into the physical part. Especially if you are using the cloud abstraction. "Administrators" manage the physical layer of things - that is why you can see
/ set the VLAN ID directly.
I believe that you would need to create your Virtual Networks with the proper VLANs and the application admins can then choose based on the Virtual Network.
The worst case is that you define a Virtual Network per VLAN so that your application admins can select an item that correlates to a specific VLAN configuration.
Brian Ehlert
http://ITProctology.blogspot.com
Learn. Apply. Repeat.

Query on Customer usage on a VMs

Hi Team,
I need assistance on the following queries
We are trying to check the feasibility of using VCloud usage meter but we are not able to understand how we can achieve following things
1) Customer X has been provided access to VM Cloud machine. some ip address machine.
2) I need to know how to configure this Customer X usage in vCloud usage meter on VM A. ITs not allowing me check the particular users usage on each VM? I mean I want to understand how this Customer X used the VM and charge him accordingly for his usage.
Like wise I have many users using the VMs.
Thanks and request you to please email me if you have any queries
Regards,
Shiva Shankar Sheela

Hello
I am facing an issue on User exit usage. I am doing following steps.
1. Dragging Calander year/month characteristic.
2. Define a varaible (Technical Name: ZCALMONTH) with Customer exit as processing by and Single value with entry "Optional" field.
3. Go in CMOD and oepn EXIT_SAPLRRS0_001.
4. Open include "ZXRSRU01".
5. Write following code and activate it.
data: l_s_range type rsr_s_rangesid.
data: l_help(8) type c.
case 'i_vnam'.
when 'ZCALMONTH'.
clear: l_s_range.
l_help = sy-datum. "actual day
l_help+6(2) = '01'. "first of month
l_help = l_help - 1. "previous month
l_s_range-low = l_help(6).
l_s_range-sign = 'I'.
l_s_range-opt = 'EQ'.
append l_s_range to e_t_range.
endcase.
After doing all above steps, variable default value remains blank. Please let me know if i am wrong or missing any step.
Thanks

Creation of Vehicle in VMS

Hi,
Pls advise the vehilce creation process in VMS.
Regards,
Student

Dear Student,
The process defined in the Best practices of VMS is as follows:
1: Select a model which has correct primary and secondary Matrices assigned, and use action CREA to create a vehicle.
2: Create purchase order
3. Create Goods Receipt
4. Incoming Invoice
5. Create sales order
6. Post PGI
7. ....
and the last is create invoice.
Once an action is performed, check the Actions list, that would help you identify the correct next step.
If you are not able to figure out, keep checking the LOG, which would give you the correct error message. If you are not able to decode the message properly, u can post it in forum here.
Hope this helps.
Best regards,
Harsha

Oracle on VMS question

Hi,
At the steel company I work at, We are starting to look at implementing a new ERP (i-Renaissance). We are seriously considering having the DB on oracle. We are thinking on having it on Windows or perhaps Unix. Some other people here are considering having it on VMS.
I am asking the broad question of the for and againsts of oracle on VMS. I am asking the question with 3 aspects in mind.
- Reliability of the DB on VMS
- Cost of the VMS implementation vs. windows. (license, OS and hardware)
- Support for potential issues (both from oracle and from community such as these forums)
What about RDB. Should we consider it?
Please keep in mind that we are at the beginning of this decision process and that I am new to oracle. I need to make an informed decision on this.
Best wishes to all!

Well thank you so far for the replies.
Currently where I work we have a bunch of windows servers for most of everything we need. Our aging ERP system is VMS/ALPHA based (the alpha is 12 years old I think). We will have the ERP company come and migrate us to the new system.
Some people here think (since that is what they know), that the DB and application server for the ERP should all be on VMS/ALPHA. I do not reject VMS outright, but I too have doubts about VMS's future although I KNOW it is a solid OS. I also wonder how long will Oracle develop the DB on VMS.
So definately we will consult with the ERP company for advice on those matters.
So my question boils down to: if running the DB and aplication on windows is a viable option for the ERP should I say no to VMS based on the cloudy future of the OS???
Also will the VMS server performance be similar to a windows server(s) solution provided we get a new alpha???
What about cost of the 2 solutions?
What about support in case of issues with the DB? Should I be worried about this on a VMS based DB?
Thanks so far for the replies, It is helping my thought process.
Later!

VMS : SQL*NET V2.0 ARCHITECTURE

제품 : SQL*NET
작성날짜 : 2001-05-28
VMS : SQL*NET V2.0 ARCHITECTURE
===============================
1. SQL*Net V2.0 Architecture
* Master File 위치 : Logical name TNS_ADMIN
= ORA_ROOT:[NETWORK.ADMIN]
* Main File 종류 : SQLNET.ORA
LISTENER.ORA
TNSNAMES.ORA
2. SQLNET.ORA
* SQL*Net에 대한 몇가지 parameter setting을 가지고 있다.
* Comments는 # 후에 기록한다.
* Client와 Server 측면에서의 trace와 log level을 setting할 수 있다.
* TRACE_FILE_xxxxx parameter는 trace file의 이름을 지정한다.
.TRC file name의 끝부분을 자동으로 생성한다.
* IPC(Inter-Process Communication)를 제어하는 parameter를 설정할 수
있는데 default는 ON이다. 이것은 자동으로 OpenVMS Mailboxe와
Unix pipes 중에서 platform에 맞게 고르는 것이다.
# FILENAME: sqlnet.ora
# NETWORK.: UK_VAX_NET
# SERVICE.: NA
TRACE_LEVEL_CLIENT = OFF
TRACE_DIRECTORY_CLIENT = ORACLE$DISK:[ORACLE7.PROD.NETWORK.TRACE]
TRACE_FILE_CLIENT = CLIENT
LOG_DIRECTORY_CLIENT = ORACLE$DISK:[ORACLE7.PROD.NETWORK.LOG]
LOG_FILE_CLIENT = CLIENT
TRACE_LEVEL_SERVER = OFF
TRACE_DIRECTORY_SERVER = ORACLE$DISK:[ORACLE7.PROD.NETWORK.TRACE]
TRACE_FILE_SERVER = SERVER
LOG_DIRECTORY_SERVER = ORACLE$DISK:[ORACLE7.PROD.NETWORK.LOG]
LOG_FILE_SERVER = SERVER
AUTOMATIC_IPC = ON
SQLNET.EXPIRE_TIME = 10
3. LISTENER.ORA
* LISTENER의 default name은 LISTENER이다.
* LISTENER의 name과 capability를 setting한다.
<lsnr-name> = (ADDRESS_LIST =
(ADDRESS =
(PROTOCOL = IPC)
(KEY = sid)
(ADDRESS =
(PROTOCOL = DECNET)
(NODE = node-name)
(OBJECT = object-name)
(ADDRESS =
(PROTOCOL = TCP)
(HOST = node-name)
(PORT = 1521)
* ORACLE8 이후로는 DECnet protocol은 지원하지 않는다.
* LISTENER START & STOP COMMAND : LISTENER 이름이 LISTENER일 경우
$ LSNRCTL START
$ LSNRCTL STOP
$ LSNRCTL STATUS
4. TNSNAMES.ORA
* Client level에서 필요한 file로 service를 define한다.
* Service Define
For DECnet...
<service-name> =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS =
(PROTOCOL = DECNET)
(NODE = <node-name>)
(OBJECT = <lsnr-object>)
(CONNECT_DATA =
(SID = <sid>)
For TCP/IP...
<service-name> =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS =
(PROTOCOL = TCP)
(HOST = <node-name>)
(PORT = 1521)
(CONNECT_DATA =
(SID = <sid>)
5. Logging and Tracing
* Listener logging은 default로 자동으로 하게 되어있다.
LOG_DIRECTORY : Log File이 생성될 Directory setting
LOG_FILE_<lis> : Log File name
* Tracing Level
TRACE_LEVEL_<listener-name> = OFF|USER|ADMIN
OFF tracing을 disable시킨다.
USER 제한적으로 tracing 한다.
ADMIN 상세하게 tracing 한다.
* server process에 대한 tracing은 default로 disable되어 있다.
6. Tips
* 실제로 동작하기 위해서 server process를 기동시켜주는 command
procedure인 ORASRV_NETV2.COM 이 있어야 하는데 TNS_ADMIN 에
sample이 있으므로 수정해서 사용하도록 한다.
이때 반드시 ORA_DB directory가 정의되어 있어야 한다.
* SYLOGIN.COM이나 LOGIN.COM에 다음의 조건문을 넣어서 Network
Access를 처리할 수 있다.
$ IF F$MODE() .EQS. "NETWORK" .OR. F$MODE() .EQS. "OTHER" -
THEN EXIT
* 만약 listener나 server process가 아무런 tracing information을
남기지 않고 계속 죽는다면 VMS Accounting Utility를 이용하여
process termination status를 점검해본다.
$ ACCOUNTING/FULL/IDENT=<pid>

Depending on how large is your database you can use either exp/imp utility to migrate the data after installing the Oracle8i on your O/S or Inplace db upgrade. Otherwise backup your current Oracle7.x release and install the Oracle8i Software only without the database. Open old database using the Oracle8i svrlmgrl utility. Upgrade using the upgrade scripts available in $ORACLE_HOME/rdbms directory. Also read the documentation for any additional information on issues on the upgrades.
I am not sure about the SQL Net versions, but SQL Net Version 2 should work against net8 listener. You could try v1 too.
Hope this helps
null

Servers/VMs per pool

Is there a best practice document out there with recommendations/hard limits for the number of servers that should be in a pool or number of vms that should be in a specific pool?
I'm trying to determine how many servers per pool I should go with as a standard in our data center.

It might be helpful in defining my question: this is the output from the Oracle VM Manager Template deployment script [note again, that this is an Oracle Server Template consisting of Oracle VM Manager to be run as a virtual machine inside Oracle Server:]
Setting up the network for Oracle VM Manager ...
Oracle VM Manager needs to be configured using a static IP address. Follow
the prompts to provide your network settings for Oracle VM Manager.
Press any key to continue...
Available network bridges:
0) xenbr0 Link encap:Ethernet HWaddr 00:1A:64:48:A8:EE
inet addr:9.123.123.123 Bcast:0.0.0.0 Mask:255.255.255.0
Please choose one of the network bridges(default is xenbr0):0
You selected the network bridge: xenbr0
Enter static IP address: 9.123.123.123
Enter netmask: [255.0.0.0] 255.255.255.0
Enter gateway: 9.23.123.123
Enter DNS server: 9.23.234.234

VMS port of BerkeleyDB

Hi,
I am wondering if you can help me with a little archaeology expedition.
I understand from the Change Log for Berkeley DB, that very early versions (i.e. up to V2.7.7) were available to be compiled on a VMS platform. If it is still possible to obtain the instructions to build the OpenVMS Alpha of Berkeley DB, this would be of immense interest to me. I've tried to apply the generic /PORT build instructions described in V1.8.5 and apply them to V1.8.5 and V2.7.7, but the builds fails, possibly due to symbolic link differences (I used VMS logicals and symbols to substitute for Unix's ln -s command) but even putting all the files in /include, etc., it still didn't compile/build. I am getting the distinct feeling that I did not grasp the full build process correct to be able to adapt it to a VMS environment.
The underlying situation is that I have Perl 5.8.6 and am migrating a Perl application from its current incarnation to this. In addition, I would like to take the opportunity to upgrade the DBM modules that it accesses to DB_File (which is the Perl interface to Berkeley DB, via yet another interface known as AnyDBM_File).
If anyone would be able to point to instructions to build Berkeley on OpenVMS Alpha 7.3-2 and specify what directory structure, source code modules, etc., that I need, it would be very beneficial experience and, hopefully (if literature is to be believed) provide good performance improvements too.
Yours gratefully,
Christopher

Hello,
One product you may want to look in to is GNV.
It is a layered product which can be installed on
some versions of VMS. It provides a Unix-like
environment with a bash shell and many utilities.
Some other information which may be useful to you is at:
http://www.oracle.com/technology/documentation/berkeley-db/db/ref/distrib/port.html
thanks,
Sandra

TNSLISTENER VS. BEQLISTENER ON VMS

제품 : ORACLE SERVER
작성날짜 : 1998-11-29
* TNSListener
TNS Listener의 기본 업무는 클라이언트의 연결 요구를 받아들여 MTS의 경우는
해당 서비스 프로토콜의 디스패처에게 연결을 넘겨주고 데디케이티드 서버
방식의 경우는 Prespawned 또는 만들어질 서버 프로세스에게 연결을 넘겨주는
역할이다. TNS Listener는 복수 개의 인스턴스에대해 서비스를 할 수 있으며,
V7.3에서 새로 생긴 Beq프로세스들은 SQL*Net 을 통하지 않으며, 이전 버젼에서
사용되던 Single-Task방식을 대체하였다.
* LSNRCTL
LSNRCTL 명령은 ORA_NETCONFIG:LSNRCTL.COM을 수행하게 되는데, 이것은 ORA_NETCONFIG:LSNRCTL.EXE파일을 수행하는 Shell역할을 하게 된다. LSNRCTL
명령은 반드시 ORACLE 계정에서 수행하여야 하며, MAXSYSGROUP값 이상의 계정
에서 수행되어야 한다.
* Listener가 갖고 있어야 할 시스템 권한
1. CMKRNL : Listener가 생성한 서버 프로세스에게 이 권한이 이양됨.
2. DETACH : Detached 프로세스를 생성할 수 있는 권한.
3. LOG_IO : I/O 함수들을 수행시킬 수 있는 권한
4. NETMBX : DECnet을 사용할 수 있는 권한.
5. TMPMBX : Temporary MailBox를 만들수 있는 권한.
6. PRMMBX : 클라이언트의 요청을 들을 수 있는 Permanent한 MailBox를 만들
수 있는 권한
7. SYSLCK : 시스템 자원에 대한 락을 걸 수 있는 권한.
8. SYSNAM : SYSTEM 로지컬 테이블과 명을 만들 수 있는 권한.
9. WORLD : 디스패처나 공유 서버 프로세스들에 대한 정보와 통제를 할 수
있는 권한.
* Listener 프로세스의 쿼타
리스너의 쿼타와 리스너가 생성하는 프로세스들에 대한 쿼타(MTS사용시의 공유
서버 프로세스는 아님)를 조절할 수 있는 방법은 로지컬 명을 이용하는 것이다.
ORA_LSNR_<쿼타 명> : 쿼타 명에 들어갈 수 있는 것들
1. ASTLM
2. BIOLM
3. BYTLM
4. CPULM
5. DIOLM
6. FILLM
7. PGFLQUOTA
8. PRCLM
9. TQELM
10. WSQUOTA
11. WSDEFAULT
12. ENQLM
13. WSEXTENT
14. JTQUOTA
위의 로지컬 명 중 일부는 ORA_NETCONFIG:LSNRCTL.COM 파일 안에 정의되어
있으며 USER 레벨로 정의되어 있는 것이 바람직하다. 만약 리스너가 많은
서비스를 핸들하고 있다면 위의 값들을 적절히 늘리는 것이 필요할 수도 있다.
리스너가 생성하는 프로세스들에 대한 쿼타를 늘리기 위해서는 리스너 프로세스
안에서 수행되는 ORA_NETWORK:TNSLSNR.COM파일을 수정하여 적절한 값을 설정
하면 된다. 기본적인 로지컬 명이 이 파일에 명시되어 있지만 디폴트는 주석처리
되어 있다. 서버 프로세스들에 대한 쿼타를 설정하는 또 다른 방법으론
LISTENER.ORA파일에 다음과 같은 OSDS 를 명시하는 방법이 있다.
이 방법은 SID별로 달리 줄 수 있다는 장점이 있다.
SID_LIST_LISTENER =
     (SID_DESC =
          (SID_NAME = ORA7)
          (PROGRAM = ORASRV_NETV2.COM)
          (OSDS =
               (PRIORITY = 8)
               (QUOTA =
                    (ASTLM =100)
                    (BYTLM=400)
                    (PGFLQUOTA=30000)
위의 OSDS에서 명시할 수 있는 쿼타의 값들은 제한이 없지만, QUOTA 리스트에
하나라도 명시되어 있는 경우에는 로지컬 명을 아무리 설정하더라도 무시되며,
QUOTA 리스트에 빠져있는 쿼타들은 시스템의 디폴트 값을 사용하게 된다.
가능하면 서버의 PRIORITY값은 늘리지 않는 것이 추천된다.
*ORASRV_NETV2.COM파일
리스너가 Prespawned 또는 일반 데디케이트 서버 프로세스를 생성할 때 수행
하는 커맨트 프로시쥬어로 다음과 같은 세개의 라인이 변경되어야 한다.
$ ora_db = "database administration directory"
$ sid = "sid name"
$ dbname = "database name"
더우기 이 ORASRV_NETV2.COM파일을 LISTENER.ORA파일에 사용하는 경우에 로지컬
명이 아닌 디바이스 및 디렉토리 명을 Full로 지정해야 한다.
* Bequeath Listener
Bequeath 리스너는 이전 버젼의 Single Task방식을 대체하는 것으로 SQL*Net
을 설정하지 않고 디폴트로 사용할 수 있는 방식이다. 이 리스너는 INSORACLE
명령이 수행되는 경우(일반적으로 시스템의 Startup에 넣어 놓음)에 자동으로
Startup이 되며, REMORACLE을 수행하기 전까지는 계속 수행된다. 만약
Bequeath 리스너가 구동되지 않은 경우에는 다음과 같은 명령으로 수동으로
Startup시킬 수 있다.
$ BEQLSNR STARTR
Bequeath 리스너는 특정 MailBox를 통하여 SQL*Net을 타지 않고 들어 오는
클라이언트의 요청(일반적으로 로컬 Connection)을 받아 들여서 Detached 서버
프로세스와 두 개의 MailBox(하나는 클라이언트의 메시지만 저장되는 MailBox,
나머지 서버가 작업 결과를 저장하는 MailBox. 이와 같이 2개로 MailBox를
나누는 이유는 클라이언트와 서버 프로세스가 서로 저장하고 읽어 가면 자신이
저장한 메시지를 자신이 읽어가는 상황이 발생할 수 있기 때문임. VMS의
MailBox는 메시지 충돌을 막아주는 기능이 없기 때문임)를 생성한 후에
클라이언트에 이 MailBox들의 이름을 전송해주게 됨으로서, 클라이언트는
서버 프로세스와 연결이 가능해 진다.
Bequeath 리스너가 클라이언트의 연결 요청을 받아들이는 데 사용되는 MailBox
는 다음과 같은 이름을 갖게 된다.
ORA_BEQ_READ_MBX_xxxxyyyyy_n
xxxx : 자동으로 생성되는 유일한 번호
yyyyy: 시스템의 이름
n : 최대 9개의 Bequeath리스너의 번호.
Bequeath 리스너를 Stop하는 명령은 DCL 프롬프트 상태에서 BEQLSNR STOP명령
으로 가능하며 Bequeath방식을 사용하는 경우에 자주 발생하는 문제와 이에
대한 해결책은 다음과 같다.
* Bequeath 방식을 사용한 경우에 발생하는 문제점의 해결책
1. Trace정보를 남기는 방법
Bequeath 리스너는 서버 프로세스를 Detached로 생성하기도 하지만 자신 역시 Detached 프로세스 방식으로 구동되기 때문에 사용자가 SYS$OUTPUT 로지컬 명을 Define하지 않으면 Trace정보를 볼 수 없다. Bequeath 리스너에 대한 Trace정보
를 보기 위해서는 다음과 같은 절차를 수행하면 된다.
a) Bequeath 리스너를 $BEQLSNR STOP 명령으로 STOP
b) STARTUP_BEQLSNR.COM파일을 EDITER로 OPEN.
c) NL: 문자열을 임의로 지정한 파일명으로 대체.
d) Bequeath 리스너 재 구동.
2. ORA-12203 : TNS: unable to connect to destination 에러가 발생하는
경우.
대부분이 Bequeath리스너가 운용중이지 않은 경우에 발생하는 에러로서,
DCL프롬프트 상태에서 BEQLSNR STATUS 명령을 사용하여 리스너가 운용중인지
를 확인한 후, 응답이 없으면 BEQLSNR STOP으로 리스너를 KILL한 후에 다시
START를 하면 됨.
3. 클라이언트에서 ORA-12203에러가 발생하는 경우.
다음과 같은 2가지 방법 중에 하나를 사용하라.
a) ORA_BEQ_TIMEOUT 로지컬 명을 120초 이상으로 (예를 들면 300초)설정
해라. 설정하는 방법은 ORA_NETWORK:BEQLSNR.COM파일에 위의 로지컬
명을 설정하는 라인을 삽입하면 된다.
b) ORA_BEQ_NUM_OF_LISTENER 로지컬 명을 정의한 후 1~10사이의 값을 세팅
해라. 이 것은 Beq 리스너 프로세스의 수를 증가시키게 되어 동시에
들어오는 클라이언트의 요청을 빨리 처리할 수 있게 된다.
* Bequeath Listener가 생성하는 서버 프로세스의 쿼타를 늘리는 방법
ORA_NETWORK:BEQLSNR.COM파일에 명시된 쿼타 파라미터의 주석처리를 풀고 적절
한 값을 세팅하면 되며, 반드시 리스너를 재구동해야 한다.

Well thank you so far for the replies.
Currently where I work we have a bunch of windows servers for most of everything we need. Our aging ERP system is VMS/ALPHA based (the alpha is 12 years old I think). We will have the ERP company come and migrate us to the new system.
Some people here think (since that is what they know), that the DB and application server for the ERP should all be on VMS/ALPHA. I do not reject VMS outright, but I too have doubts about VMS's future although I KNOW it is a solid OS. I also wonder how long will Oracle develop the DB on VMS.
So definately we will consult with the ERP company for advice on those matters.
So my question boils down to: if running the DB and aplication on windows is a viable option for the ERP should I say no to VMS based on the cloudy future of the OS???
Also will the VMS server performance be similar to a windows server(s) solution provided we get a new alpha???
What about cost of the 2 solutions?
What about support in case of issues with the DB? Should I be worried about this on a VMS based DB?
Thanks so far for the replies, It is helping my thought process.
Later!

Libvirt: Unable to define LVM storage pool

Hello,
I'm trying to define an LVM storage pool for my virtual machines using KVM/libvirt. The configuration looks like this:
<pool type="logical">
<name>vol0</name>
<source>
<device path="/dev/md0"/>
</source>
<target>
<path>/dev/vol0</path>
</target>
</pool>
The problem is, that this LVM group is already active (other vms running using volumes inside this group) and 'virsh pool-start vol0' wants me to disable it. Is there any way to start the pool without "deactivate" the volume group?
virsh pool-start vol0
error: internal error '/sbin/vgchange -an vol0' exited with non-zero status 5 and signal 0: Can't deactivate volume group "vol0" with 14 open logical volume(s)
Further, I'm a bit curious that libvirt might recreate the volume group and therefore deletes all the content during the building process.
Would appreciate any advice
Regards,
Jonas

maahes wrote:
did so now, only now I'm getting a slightly different error: could not find udevd no such file or directory. I checked both grub.cfg's and my mkinitcpio.conf and there's no listing for udevd ....which I've never heard of, so I assumed it was a typo?
For clarification: udev is in the mkinitcpio.
I'm not sure whether I yet have a good intuition for how you have your machine set up, but I suspect you need to include a cryptdevice flag to the kernel in your grub config. The file isn't found because the kernel doesn't know your root directory needs decrypting first.
My setup is an LVM over LUKS over LVM sandwich. To boot into my system, the grub.cfg contains the line:
linux /vmlinuz-linux root=/dev/mapper/cryptvg-root cryptdevice=/dev/mapper/vg-crypt:root rootfstype=ext4 pcie_aspm=force acpi_osi=Linux acpi_backlight=vendor i915.i915_enable_rc6=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1 ro
Now, most of those flags don't have anything to do with your problem, but note the cryptdevice. It tells the kernel it's dealing with an encrypted filesystem sitting in a logical volume called crypt on a volume group called vg. The bit after the colon tells the kernel to associate this encrypted filesystem with /dev/mapper/root.
As for how to fix your system, I'm afraid I still feel a bit fuzzy about how your LVM and encrypted layers relate to each other, whether you have LVM over LUKS, or LUKS over LVM, or something else. Was there a particular how-to that you followed?

Oracle VM - defining a static network

Hi,
I'm not sure if I'm thinking about this right or not, but I'm trying to define an interface on a VM that uses a static ip address. I'm doing this through the web services API, but I would be happy to better understand how OVM does networking.
In any case, I'm calling the addNIC command which requires a VM and a NICconfig. The NICconfig has a place for IP address, netmask, etc. Here is an example set of args I provide:
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<jbi:message xmlns:msgns="http://oracle.ovs.api/" type="msgns:LifecycleService_addNIC" version="1.0" xmlns:jbi="http://java.sun.com/xml/ns/jbi/wsdl-11-wrapper">
<jbi:part>
<addNICElement xmlns="http://oracle.ovs.api/types/">
<virtualMachine xmlns:ns0="http://oracle.ovs.api/types/" xmlns:ns1="http://www.oracle.com/webservices/internal/literal" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:type="ns0:VirtualMachine">
<ns0:imgName>test-vm-2</ns0:imgName>
<ns0:comments xsi:nil="1"/>
<ns0:imgId>110</ns0:imgId>
<ns0:bootDevice>
<ns0:bootDevice>HDD</ns0:bootDevice>
</ns0:bootDevice>
<ns0:imgSize>6205</ns0:imgSize>
<ns0:vncPassword>XXXXXXXX</ns0:vncPassword>
<ns0:maximumMemory>1024</ns0:maximumMemory>
<ns0:publish xsi:nil="1"/>
<ns0:cpuNumber>1</ns0:cpuNumber>
<ns0:haEnable>false</ns0:haEnable>
<ns0:uuid>69949450-d3b9-4ffa-a0a2-8b0b7c4198c7</ns0:uuid>
<ns0:operatingSystemName>Oracle Enterprise Linux 5 64-bit</ns0:operatingSystemName>
<ns0:createTime>1257357624422</ns0:createTime>
<ns0:parentImg>90</ns0:parentImg>
<ns0:systemPassword xsi:nil="1"/>
<ns0:location>110_test-vm-2</ns0:location>
<ns0:memory>1024</ns0:memory>
<ns0:systemLogin xsi:nil="1"/>
<ns0:platform xsi:nil="1"/>
<ns0:deploy>Auto</ns0:deploy>
<ns0:status>
<ns0:code>2</ns0:code>
<ns0:status>Powered Off</ns0:status>
<ns0:vmId>0</ns0:vmId>
</ns0:status>
<ns0:vncPort xsi:nil="1"/>
<ns0:imgIp xsi:nil="1"/>
<ns0:poweronTime xsi:nil="1"/>
<ns0:architecture xsi:nil="1"/>
<ns0:virtualMachineType>
<ns0:vmType>Paravirtualized</ns0:vmType>
</ns0:virtualMachineType>
<ns0:keyboardLayout>en-us</ns0:keyboardLayout>
<ns0:pvdriver>false</ns0:pvdriver>
<ns0:cpuCap>100</ns0:cpuCap>
<ns0:cpuPriority>50</ns0:cpuPriority>
</virtualMachine>
<nicConfig>
<ipAddress>129.146.137.195</ipAddress>
<subnetMask>255.255.255.0</subnetMask>
<qosEnabled>false</qosEnabled>
<dnsServer/>
<frontEndName/>
<domainName/>
<defaultGateway/>
<hostName>vm-test2</hostName>
<rateLimit>0.0</rateLimit>
<bridge>xenbr0</bridge>
<networkConfType>
<type>STATIC</type>
</networkConfType>
<name>VIF0</name>
<networkType>
<networkType>NETFRONT</networkType>
</networkType>
<macAddress>00:16:3E:11:11:11</macAddress>
</nicConfig>
</addNICElement>
</jbi:part>
</jbi:message>
However, when I do an ifconfig -a on the VM, there are two main issues:
1. The eth0 interface looks to be preconfigured, but does not show up when I query the interfaces on the VM
2. The eth1 has the MAC address provided above, but does not have the IP address or netmask set
Some side questions:
* what are appropriate values for "frontEndName"?
* can Oracle VM generate a MAC address for me using the API? If so, how? I tried to leave it empty and got an error indicating that it must be provided.
Thanks in advance,
John

jxstanford wrote:
That makes sense (not that it's proper). I'm basically seeing what you described which is an initial interface that shows up in vm.cfg, but not in the OVM GUI. I'm still struggling to define a static interface with an IP address. When you mass create your VMs, are you using DHCP to assign addresses, or do you have a solution for setting up a static IP address?I have a solution: I use a source text file that contains data like server hostname, IP address, netmask, static routes, etc. I then cycle through each VM and loop-mount the system.img itself. Once mounted, I inject the configuration details into the appropriate files on the disk (like a ninja!). This is possible for both normally partitioned and LVM-based virtual disks, though the LVM stuff requires a little bit more work to enable/disable the volume groups each time. If you ping me at my [email protected], I can send you the (very dodgy) script I've been using so far.
Note however that this is something that Enterprise Manager's VM Management Pack can do in a GUI environment: You are able to mass create/configure guests using EM and it will use the same mechanism to inject IP configuration data into the guest prior to boot. It can also auto-configure EM agent as well.

AttachCurrentThread fails when several VMs running in different threads

Hi,
I would like to build an application that runs several JVMs in parallel. I'm using Java6 and according to the JNI specs this should be possible as long as each VM instance is running in a separate thread.
In the below mentioned demo application two threads are created. Both instantiate a VM (which works well). But only one thread seems to be able to run AttachCurrentThread. When AttachCurrentThread is called for the second one I get the following error:
# An unexpected error has been detected by Java Runtime Environment:
# SIGSEGV (0xb) at pc=0x080486ae, pid=1428, tid=3035118512
# Java VM: Java HotSpot(TM) Client VM (1.6.0_03-b05 mixed mode, sharing)
# Problematic frame:
# C [jvmthreads+0x6ae]But because one thread successfully instantiates a java class I think that the basic code structure is OK. It seems to be a threading issue. My application looks as follows:
#include <stdlib.h>
#include <jni.h>
#include <pthread.h>
#define USER_CLASSPATH "myclasses.jar"
#define CLASS_NAME "com.foobar/DummyRtbosFiller"
JavaVM* JavaThreads_createJVM(){
     JNIEnv *env;
    JavaVM *vm;
    jint res;
    setenv("JAVA_VM_VERSION", "1.5", 1);
//    setenv("JAVA_VM_VERSION", "1.5", 0);
    JavaVMInitArgs vm_args;
    JavaVMOption options[7];
    options[0].optionString = malloc(3000*sizeof(char));
    sprintf(options[0].optionString, "-Djava.class.path="USER_CLASSPATH);
    options[1].optionString = "-Xmx64m";
     int enableRemDebugging = 0;
     if(enableRemDebugging){
         printf("enable remote debugging");
          // intellij remote debugging support
          options[2].optionString = "-Xdebug";
          options[3].optionString = "-Xnoagent";
          options[4].optionString = "-Djava.compiler=NONE";
          options[5].optionString = "-Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=34343";
          vm_args.nOptions = 6;
     }else{
          vm_args.nOptions = 2;
     vm_args.options = options;
    vm_args.version = JNI_VERSION_1_4;
    vm_args.ignoreUnrecognized = JNI_TRUE;
    /* Create the Java VM */
    res = JNI_CreateJavaVM(&vm, (void**)&env, &vm_args);
    printf("java started.\n");
     return vm;
* \brief Shuts down an existing java virtual machine.
* \param jvm Pointer to the jvm to be destroyed.
* This method is used to shut down an existing java virtual machine.
int JavaThreads_destroyJVM(JavaVM *jvm){
//     JNIEnv *env = getAttachedEnv(self, jvm);
//     if ((*env)->ExceptionOccurred(env)) {
//        (*env)->ExceptionDescribe(env);
      // detach the current thread from the vm
    (*jvm)->DetachCurrentThread(jvm);
    (*jvm)->DestroyJavaVM(jvm);
     return 0;
* \brief Attaches the current thread to a given vm instance
* \param self Pointer to bbcm-instance
* \param jvm Pointer to the jvm use for attachment
* This method is used to attach the current thread to a given vm instance
JNIEnv* getAttachedEnv(JavaVM *jvm){
     JNIEnv *localEnv = NULL;
     int envErr = 0;
     /* get a local java env */
     envErr = (*jvm)->AttachCurrentThread( jvm, (void**)&localEnv, NULL );
     if ( envErr != 0 ){
          if ( (*localEnv)->ExceptionCheck( localEnv ) == JNI_TRUE ){
            (*localEnv)->ExceptionDescribe( localEnv );
     if((*localEnv)->ExceptionOccurred(localEnv)){
         (*localEnv)->ExceptionDescribe(localEnv);
     if (localEnv == NULL) {
        printf("ERROR: failed to get ENV pointer in pushContext");
          // JavaThreads_destroyJVM(jvm);
        return (JNIEnv*) NULL;
     return localEnv;
* \brief Create a new instance of the transcriber application
* \return Pointer to the java object
* This method is used to create a new instance transcriber application
jobject startJavaSubSystem(JavaVM *jvm, char* argString){
    printf("attaching env pointer.\n");
     JNIEnv *env = getAttachedEnv(jvm);
    printf("thread attachment done.\n");
     jclass cls;
     jmethodID mid;
     jobject transcriber;
     cls = (*env)->FindClass(env, CLASS_NAME);
    if (cls == NULL) {
        JavaThreads_destroyJVM(jvm);
     jstring jArgumentString = (*env)->NewStringUTF(env, argString);
    mid = (*env)->GetMethodID(env, cls, "<init>", "(ILjava/lang/String;)V");
     if (mid == NULL) {
        JavaThreads_destroyJVM(jvm);
    printf("starting object instantiation\n");
    (*env)->NewObject(env, cls, mid, 1, jArgumentString);
     return transcriber;
void startJavaThread(char* arg){
     JavaVM *jvm;
    printf("create a jvm instance for component '%s'\n", arg);
     jvm = JavaThreads_createJVM();
    printf("attempting to start component '%s'\n", arg);
    startJavaSubSystem(jvm, arg);
    printf("component '%s' started !\n", arg);
    // sleep forever to keep the thread alive
    sleep(10000);
int main(int argc, char **argv) {
      pthread_t thread1, thread2;
      /* Create independent threads each of which will instantiate a java VM. */
     int iret1 = pthread_create( &thread1, NULL, startJavaThread, (void*) "test47");
     sleep(1);
     int iret2 = pthread_create( &thread2, NULL, startJavaThread, (void*) "test48");
     pthread_join( thread1, NULL);
     pthread_join( thread2, NULL);
     printf("Thread 1 returns: %d\n",iret1);
     printf("Thread 2 returns: %d\n",iret2);
     return 0;
}What could be my mistake?
Any help (links, docs, ideas) is welcome.
Best regards, Holger

You're right. I got confused by the specs of JNI_GetCreatedJVMs:
JNI_GetCreatedJavaVMs function returns all virtual machine instances
that have been created in the current process.But unfortunately you're completely right (cf. http://java.sun.com/javase/6/docs/technotes/guides/jni/spec/invocation.html#wp9502 )
As of JDK/JRE 1.2, creation of multiple VMs in a single process is not supported.Nothing seems to have changed since then. :-(
thanks for the help, bg, holger

Defining logicals on VMS

Similar Messages

Maybe you are looking for