Re: Shared and non-shared ques on a single que manager?
Ed, Shared Queue messages reside in a Coupling Facility and therefor can't be bigegr than 63(64 - 1)K . This has nothing to do with using one or more Qmanagers. Regards, Erik Dijkerman X Rabobank ICT/Serverbedrijf PIM/OS390 ZL-S206400 Mailbox 17100, 3500 HG Utrecht *(030) 215 4878 *(030) 215 3085 ? [EMAIL PROTECTED] Somewhere, something incredible is waiting to be known. --Blaise Pascal -Original Message- From: Perfetto, Ed [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 9:27 PM To: [EMAIL PROTECTED] Subject: Shared and non-shared ques on a single que manager? New subscriber here!! Can you or can you not have shared and non-shared ques on a single que manager? We're trying to overcome the 63K limits in a sysplex environment and have decided to have multiple que managers, some shared(<63K) and others not(>63K). Thanks, Ed. De informatie opgenomen in dit bericht kan vertrouwelijk zijn en is uitsluitend bestemd voor de geadresseerde. Indien u dit bericht onterecht ontvangt, wordt u verzocht de inhoud niet te gebruiken en de afzender direct te informeren door het bericht te retourneren. The information contained in this message may be confidential and is intended to be exclusively for the addressee. Should you receive this message unintentionally, please do not use the contents herein and notify the sender immediately by return e-mail. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
Hi Rick Thankyou for thoughts - I think that could be part of the explanation. Though if the queue manager fires one Trigger first then the arrival of another message on the queue will not trigger the condition for trigger first (that the depth goes from 0 to 1) as the depth is already > 0 and nothing has taken a message off. Unless it is the fact that during the longer CICS start up time the first message has expired and something has caused the queue manager to reset the queue depth to zero, so it fires off another when the next message arrives as IPROCS on the queue is still zero. My transaction has a get with WAIT for the next message to arrive so it does not shut down immediately when there is no message to process. Messages arrive every second or so for it to process. Alan Alan Turnbull Senior Developer QBEMM Direct: (02) 8275 9880 [EMAIL PROTECTED] |+--> || Rick Tsujimoto | || <[EMAIL PROTECTED]| || CANON.COM> | || Sent by: MQSeries List | || <[EMAIL PROTECTED]> | || | || | || 04/04/2003 10:03 AM | || Please respond to MQSeries | || List| || | |+--> >--| | | | To: [EMAIL PROTECTED] | | cc: | | Subject: Re: Triggering Z/OS DLQH CSQUDLQH | >--| One possibility could be a timing issue. During a cold start, more processing could be occuring vs. a warm start. Since CKTI is fired from the PLTPI, it could have started the first instance of your triggered transaction. If, for some reason, that instance got hung up, e.g. busy system, and didn't have a chance to open the queue, another trigger message could be created if another message arrives in the interim, causing the second instance to be created My question is, why doesn't your transaction terminate if it finds nothing on the queue? [EMAIL PROTECTED] EMM.COM.AU To: [EMAIL PROTECTED] Sent by: cc: MQSeries ListSubject: Re: Triggering Z/OS DLQH CSQUDLQH <[EMAIL PROTECTED] en.AC.AT> 04/03/2003 05:39 PM Please respond to MQSeries List Hi I have once seen this abberrant behaviour of two trigger FIRST firing in CICS - i wonder if anyone can explain why? My CICS transaction is started by TRIGGER=FIRST on one queue and once it is started it stays alive processing messages that arrive on the queue. Often when CICS starts up there will messages on the queue and only one instance of the transaction is started as expected. One day the development CICS region crashed (cancelled by ops) and after a cold restart I later saw that I had two instances of my trigger=first transaction running. Both started at virually the same time when CICS came up - nearly consecutive transaction numbers. Both were processing the same queue as the transaction reports what queue it is started for, and nothing else is taking messages off the queue. There would have been expired and expiring messages on the queue, so could this have caused the queue manager to create a additional trigger message? This is CICS 4.1, MQ 5.2 on OS390 2.6 Thanks Alan Alan Turnbull Senior Developer QBEMM Direct: (02) 8275 9880 [EMAIL PROTECTED] Robert Broderick <[EMAIL PROTECTED] To: [EMAIL PROTECTED] TMAIL.COM>cc: Sent by: MQSeries Subject: Re: Triggering Z/OS DLQH CSQUDLQH List <[EMAIL PROTECTED] AC.AT> 04/04/2003 07:21 AM Please respond to MQSeries List The transaction will not trigger. One of the rules for first is that the queue is not opened for input. >From: Ronald Weinger <[EMAIL PROTECTED]> >Reply-To: MQSeries List <[EMAIL PROTECTED]> >To: [EMAIL PROTECTED] >Subject: Re: Triggering Z/OS DLQH CSQUDL
Re: Authorizations in 5.3
Ralph, See the following code fragment "strncpy(odrequest.ObjectName, req, (size_t)MQ_Q_NAME_LENGTH); <- Supplying queue name "DONT.ALLOW.CONNECTIONS" strncpy(odrequest.ObjectQMgrName,"MQDEV",48); O_options = MQOO_OUTPUT + MQOO_FAIL_IF_QUIESCING; MQOPEN(Hcon, /* connection handle */ &odrequest, /* object descriptor for queue */ O_options, /* open options */ &Hobj_req, /* object handle */ &OpenCode, /* MQOPEN completion code */ &Reason); /* reason code */ /* printf("Return code from openqueues: Reason = %d\n",Reason); */ /* report reason, if any; stop if failed */ if (Reason != MQRC_NONE) {" This program issues an MQCONN to a queue manager called MQTEST (which has channel sdr/rcvr connections to MQDEV) I then use the code fragment above to create an MQOPEN call on a queue called "DONT.ALLOW.CONNECTIONS" which is a local queue on MQDEV. In other words, there is no QREMOTE definition to DONT.ALLOW.CONNECTIONS on MQDEV, the Qname is being fully qualified by the addition of the QmgrName in the OPEN The call succeeds and I can use the returned objecthandle to put messages on to the queue DONT.ALLOW.CONNECTIONS on queue manager MQDEV. If I go to the platform that hosts MQDEV and do an MQCONN (to MQDEV) followed by an MQOPEN, it fails with a 2035 (which it should). I think the queue insertion from MQTEST is happening because the channel connection from MQTEST to MQDEV is running with an mca of "mqm" and therefore succeeds George Beaton DB&MW 62448125 spectrum 148125 "Jones, Ralph C {PBSG}" <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 04/04/2003 04:50 Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: Authorizations in 5.3 George, I am not sure that I follow this example. Applications should only need authorization on the QREMOTE or QALIAS queues. There is no need to authorize applications to the transmission queue. -Original Message- From: George Beaton [mailto:[EMAIL PROTECTED] Sent: Wednesday, April 02, 2003 5:57 PM To: [EMAIL PROTECTED] Subject: Re: Authorizations in 5.3 There's another little problem with it that I overlooked. If you connect to a queue manager other than the one over which the security profiles are implemented then you can issue an MQOPEN using ObjectName ObjectQmanagerName and follow up with a series of MQPUTs to the queue. Consequently you need to put security onto every transmission queue to the secured Queue manager George Beaton DB&MW 62448125 spectrum 148125 "Jones, Ralph C {PBSG}" <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 03/04/2003 08:37 Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: Authorizations in 5.3 The REFRESH SECURITY command was introduced with 5.2 on the distributed code base. The behavior concerning authorization being granted to all members of a security group when issued to a member only occurs in UNIX environments. -Original Message- From: Neil Casey [mailto:[EMAIL PROTECTED] Sent: Tuesday, April 01, 2003 5:54 PM To: [EMAIL PROTECTED] Subject: Re: Authorizations in 5.3 Hi Dave, REFRESH SECURITY was not available to unix or windows systems until 5.3. It has always been implemented on zOS and was used to refresh the RACF profiles in storage. The security capabilities of 5.3 have been extended quite a bit. So much so that there is now a book dedicated to just that topic. You might want to review it, especially looking at generic queue names in security profiles, as well as the REFRESH requirements. Another thing that can be confusing about the security cache is that it does not get preloaded. I believe that security details for an object are only loaded when the object is referenced by the particular user. This means that if you define the security before the object is referenced, it will work without a refresh, but once the user has attempted to access the object and the access failed, you need to run REFRESH, or restart the queue manager, to get it to work. You can look at all your current profiles with amqoamd or dmpmqaut (they return the same info in different formats). Also, be careful of defining new objects using a user other than mqm. If the primary group of user 'neil' is 'staff', and user 'neil' is a member of 'mqm', then 'neil' can use runmqsc and define objects. However, the 'staff' group is automatically given +all access to those objects. This is not a bug because it is documented to work this way in the manuals. It just isn't very sensible in my opinion. Is it possible that this behaviour is causing the confusion in results that you are seeing? Rgeards, Neil Ca
WMQ 5.3 on Linux RedHat 7.3
Hi all, I've just recently installed WMQ 5.3 on Linux for the first time:- weusually run WMQ on AIX. I'm surprised the see the large number of WMQ processes under Linux compared to those under AIX. (The display below is from pstree(1) because it's compact): Under AIX: |--+= 88508 mqm amqzxma0 -m MQCNDEV3 | |--- 29062 mqm amqharmx MQCNDEV3 /var/mqm | |--- 42980 mqm amqzllp0 -mMQCNDEV3 ? | |--- 47814 mqm amqzlaa0 -mMQCNDEV3 -fip2 | |--- 54084 mqm /usr/mqm/bin/amqzdmaa -m MQCNDEV3 | |--- 68368 mqm amqzlaa0 -mMQCNDEV3 -fip1 | |--- 71718 mqm amqhasmx MQCNDEV3 /var/mqm | |--- 117848 mqm /usr/mqm/bin/runmqchi -m MQCNDEV3 | |--- 125586 mqm amqzlaa0 -mMQCNDEV3 -fip0 | |--- 137190 mqm /usr/mqm/bin/amqzfuma -m MQCNDEV3 | \--- 148198 mqm /usr/mqm/bin/amqrrmfa -t2332800 -s2592000 -p2592000 -g5184000 -c3600 -m MQCNDEV3 whereas under Linux, a similar setup yeilds: init-+-amqpcsea---amqpcsea---amqpcsea |-amqzxma0-+-amqhasmx---amqhasmx---amqhasmx | |-amqrrmfa---amqrrmfa-+-amqrrmfa | | `-amqrrmfa | |-amqzdmaa---amqzdmaa-+-amqzdmaa | | `-amqzdmaa | |-amqzfuma---amqzfuma-+-amqzfuma | | `-amqzfuma | |-amqzlaa0---amqzlaa0-+-amqzlaa0 | | |-amqzlaa0 | | |-amqzlaa0 | | |-amqzlaa0 | | |-amqzlaa0 | | |-amqzlaa0 | | `-amqzlaa0 | |-amqzllp0---amqzllp0-+-amqzllp0 | | `-amqzllp0 | |-amqzxma0-+-amqzxma0 | | |-amqzxma0 | | `-amqzxma0 | `-runmqchi---runmqchi-+-runmqchi ||-runmqchi |`-runmqchi Is this normal ? Can some explain what's going on here ? Regards, Andrew Gardner E-Business Enablement Services IBM Global Services Australia +61 2 61242151 (Ph) +61 2 61242155 (Fx) +61 41 7911179 (Mb) Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
One possibility could be a timing issue. During a cold start, more processing could be occuring vs. a warm start. Since CKTI is fired from the PLTPI, it could have started the first instance of your triggered transaction. If, for some reason, that instance got hung up, e.g. busy system, and didn't have a chance to open the queue, another trigger message could be created if another message arrives in the interim, causing the second instance to be created My question is, why doesn't your transaction terminate if it finds nothing on the queue? [EMAIL PROTECTED] EMM.COM.AU To: [EMAIL PROTECTED] Sent by: cc: MQSeries ListSubject: Re: Triggering Z/OS DLQH CSQUDLQH <[EMAIL PROTECTED] en.AC.AT> 04/03/2003 05:39 PM Please respond to MQSeries List Hi I have once seen this abberrant behaviour of two trigger FIRST firing in CICS - i wonder if anyone can explain why? My CICS transaction is started by TRIGGER=FIRST on one queue and once it is started it stays alive processing messages that arrive on the queue. Often when CICS starts up there will messages on the queue and only one instance of the transaction is started as expected. One day the development CICS region crashed (cancelled by ops) and after a cold restart I later saw that I had two instances of my trigger=first transaction running. Both started at virually the same time when CICS came up - nearly consecutive transaction numbers. Both were processing the same queue as the transaction reports what queue it is started for, and nothing else is taking messages off the queue. There would have been expired and expiring messages on the queue, so could this have caused the queue manager to create a additional trigger message? This is CICS 4.1, MQ 5.2 on OS390 2.6 Thanks Alan Alan Turnbull Senior Developer QBEMM Direct: (02) 8275 9880 [EMAIL PROTECTED] Robert Broderick <[EMAIL PROTECTED] To: [EMAIL PROTECTED] TMAIL.COM>cc: Sent by: MQSeries Subject: Re: Triggering Z/OS DLQH CSQUDLQH List <[EMAIL PROTECTED] AC.AT> 04/04/2003 07:21 AM Please respond to MQSeries List The transaction will not trigger. One of the rules for first is that the queue is not opened for input. >From: Ronald Weinger <[EMAIL PROTECTED]> >Reply-To: MQSeries List <[EMAIL PROTECTED]> >To: [EMAIL PROTECTED] >Subject: Re: Triggering Z/OS DLQH CSQUDLQH >Date: Thu, 3 Apr 2003 15:54:29 -0500 > >If a CICS transaction, for example, starts due to trigger on first, and >then does a GET WAIT UNLIMITED, it is possible than a new message on the >queue will trigger a second execution of the transaction, but the current >execution will retrieve the message before the second one is attached. >So the second will get a 'no-message-available', and if coded to issue a >GET WAIT UNLIMITED, will remain suspended. A third message will trigger >another transaction, but it will be retrieved by one of the two currentl;y >suspended transactions. That sequence could continue until max tasks are >reached. > >However, have you found a MVS batch trigger monitor ? > > > > > > "Gary P. Klos" > <[EMAIL PROTECTED]> To: >[EMAIL PROTECTED] > Sent by: cc: > "MQSeries List" Subject: Triggering Z/OS >DLQH CSQUDLQH > <[EMAIL PROTECTED] > n.AC.AT> > > > 04/03/2003 02:59 > PM > Please respond to > "MQSeries List" > > > > > > > > I am setting up to use the Z/OS utility CSQUDLQH, which handles the > dead letter queue for a queue manager. What I was thinking was to put > a trigger on FIRST on the dead letter queue and once it starts > CSQUDLQH, then I could leave it running indefinitely. My question is > this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH > removes all messages from the queue and stays running, then later > another message goes on the dead letter queue, CSQUDLQH will start > again. Does that mean I will have two batch jobs running CSQUDLQH, or > does this program know to not run if it is already running with the > same dead letter queue? > > Thanks, > Gary > >Instructions for managing your mailing list subscription are provided in >the Listserv General Users Guide available at http://www.lsoft.com >Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > > > > > >The infor
SSL and channel exits (was Compatibility of Channel Encryption Methods)
Based on the answer to question 2 in the attached note, i.e., that channel "exits do not get called until after the SSL handshake has taken place," is it safe to assume that any data passed through the channel on behalf of the channel exits would be encrypted? (This is assuming the SSLCIPH is coded to use a cipherspec that provides encryption.) For example, if a SVRCONN channel is defined with SSLCAUTH set to OPTIONAL so that the clients to not need to provide a certificate, SSL could be used to encrypt the client channel, but a security exit could be used to pass a userid and password for authentication.In such a case, could we rely on SSL to encrypt values passed for the security exit, such as the password? Hoping someone from development can provide an answer to this. thanks, Tom == Tom Schneider / IBM Global Services (513) 533-3644 [EMAIL PROTECTED] == Morag Hughson <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 01/27/2003 06:59 AM Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject:Re: Compatibility of Channel Encryption Methods? Bill, A1) The code to run SSL channels (i.e. WebSphere MQ V5.3) must be available at both end of the channel in order for the channel to be able to use SSL. So, yes, both queue managers must be at V5.3. A2) You cannot write a channel exit to communicate with the SSL code in V5.3 since the exits do not get called until after the SSL handshake has taken place. So unless you have V5.3 at both ends, the channel will fail to start if SSL is specified on only one end, and you will never get as far as calling the exits. A3) As I understand it, MQSecure is written using Channel Exits, so the answer to Q2) applies. You can use V5.3 SSL on both ends to do the handshake and specify no encryption on the channel using one of the NULL_* CipherSpecs, then what you do in an exit to the data flowed is up to you. So they can interact in this way. Hope this helps Cheers Morag Morag Hughson WebSphere MQ for z/OS Development Internet: [EMAIL PROTECTED] Bill A Lee/CanWest/[EMAIL PROTECTED]To: [EMAIL PROTECTED] BMCA cc: Sent by: MQSeriesSubject: Compatibility of Channel Encryption Methods? List <[EMAIL PROTECTED] N.AC.AT> 24/01/2003 18:19 Please respond to MQSeries List Company ABC will be connecting their OS/390 queue managers (yet to be installed) to those of their customer DEF via SSL encrypted channels, PKI digital certificates, etc. DEF is their first customer to connect in this manner, but others are anticipated. Some of the ways to do this are: 1. Install WebSphere MQ for OS/390 v5.3 and its SSL channels. 2. Write custom channel exits using RSA Security's BSAFE toolkit. 3. Install a product like Candle's MQSecure (latest version is now called PathWAI Secure for WebSphere MQ) and use the channel exits it provides. What are the compatibility issues for these encrypted channels? 1. If ABC is running v5.3 and wants to use its SSL channels, is it mandatory that DEF also run v5.3 to be compatible? 2. Can custom channel exits be written by ABC using the BSAFE toolkit so they will be compatible with another encryption method, such as the SSL channels in v5.3, or in MQSecure? 3. Can MQSecure be configured to be compatible with v5.3's SSL channels, etc.? In general, ABC is hoping to avoid installing and supporting a different encryption method for each of their connected customers. Is there a way to do this? All responses are much appreciated! Thanks, ..Bill.. Heisenberg may have slept here. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
Hi I have once seen this abberrant behaviour of two trigger FIRST firing in CICS - i wonder if anyone can explain why? My CICS transaction is started by TRIGGER=FIRST on one queue and once it is started it stays alive processing messages that arrive on the queue. Often when CICS starts up there will messages on the queue and only one instance of the transaction is started as expected. One day the development CICS region crashed (cancelled by ops) and after a cold restart I later saw that I had two instances of my trigger=first transaction running. Both started at virually the same time when CICS came up - nearly consecutive transaction numbers. Both were processing the same queue as the transaction reports what queue it is started for, and nothing else is taking messages off the queue. There would have been expired and expiring messages on the queue, so could this have caused the queue manager to create a additional trigger message? This is CICS 4.1, MQ 5.2 on OS390 2.6 Thanks Alan Alan Turnbull Senior Developer QBEMM Direct: (02) 8275 9880 [EMAIL PROTECTED] Robert Broderick <[EMAIL PROTECTED] To: [EMAIL PROTECTED] TMAIL.COM>cc: Sent by: MQSeries Subject: Re: Triggering Z/OS DLQH CSQUDLQH List <[EMAIL PROTECTED] AC.AT> 04/04/2003 07:21 AM Please respond to MQSeries List The transaction will not trigger. One of the rules for first is that the queue is not opened for input. >From: Ronald Weinger <[EMAIL PROTECTED]> >Reply-To: MQSeries List <[EMAIL PROTECTED]> >To: [EMAIL PROTECTED] >Subject: Re: Triggering Z/OS DLQH CSQUDLQH >Date: Thu, 3 Apr 2003 15:54:29 -0500 > >If a CICS transaction, for example, starts due to trigger on first, and >then does a GET WAIT UNLIMITED, it is possible than a new message on the >queue will trigger a second execution of the transaction, but the current >execution will retrieve the message before the second one is attached. >So the second will get a 'no-message-available', and if coded to issue a >GET WAIT UNLIMITED, will remain suspended. A third message will trigger >another transaction, but it will be retrieved by one of the two currentl;y >suspended transactions. That sequence could continue until max tasks are >reached. > >However, have you found a MVS batch trigger monitor ? > > > > > > "Gary P. Klos" > <[EMAIL PROTECTED]> To: >[EMAIL PROTECTED] > Sent by: cc: > "MQSeries List" Subject: Triggering Z/OS >DLQH CSQUDLQH > <[EMAIL PROTECTED] > n.AC.AT> > > > 04/03/2003 02:59 > PM > Please respond to > "MQSeries List" > > > > > > > > I am setting up to use the Z/OS utility CSQUDLQH, which handles the > dead letter queue for a queue manager. What I was thinking was to put > a trigger on FIRST on the dead letter queue and once it starts > CSQUDLQH, then I could leave it running indefinitely. My question is > this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH > removes all messages from the queue and stays running, then later > another message goes on the dead letter queue, CSQUDLQH will start > again. Does that mean I will have two batch jobs running CSQUDLQH, or > does this program know to not run if it is already running with the > same dead letter queue? > > Thanks, > Gary > >Instructions for managing your mailing list subscription are provided in >the Listserv General Users Guide available at http://www.lsoft.com >Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > > > > > >The information contained in this message may be CONFIDENTIAL and is for >the intended addressee only. Any unauthorized use, dissemination of the >information, or copying of this message is prohibited. If you are not the >intended addressee, please notify the sender immediately and delete this >message. > >Instructions for managing your mailing list subscription are provided in >the Listserv General Users Guide available at http://www.lsoft.com >Archive: http://vm.akh-wien.ac.at/MQSeries.archive _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive IMPORTANT NOTICE : The information in this email is confidential and may also be pri
Re: Shared and non-shared ques on a single que manager?
Personally I'd go with one queue manager per lpar requiring MQ's facilities, supporting both shared and non-shared queues. Aside from the 63k upper limit on messages I'd be looking at expected message lifetime as the other major criteria for deciding on whether or not a queue should be shared or not. Given the finite resources in your coupling facility I'd start slowly picking the best candidates for sharing a couple at a time. Select your highest throughput request/reply pairs of queues and test thoroughly in your Test/Dev lpars before moving to production also gathering response time statistics and any decrease in utilisation of TCPIP, DASD subsystem etc.. If queue sharing is truly effective for you then you can consider upgrading your coupling facility. Regards Tim A "Perfetto, Ed" <[EMAIL PROTECTED]To: [EMAIL PROTECTED] INC.COM> cc: Sent by: MQSeriesSubject: Shared and non-shared ques on a single que manager? List <[EMAIL PROTECTED] N.AC.AT> 04/04/2003 05:26 Please respond to MQSeries List New subscriber here!! Can you or can you not have shared and non-shared ques on a single que manager? We're trying to overcome the 63K limits in a sysplex environment and have decided to have multiple que managers, some shared(<63K) and others not(>63K). Thanks, Ed. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Log Files.
Bobbee Thank you yes the product is ReQuest from Cressida. Someone was kind enough to contact me of-list with all the info. Here is a link for those interested: http://www.cressida.info/products_am_mqla.shtml Ian Robert Broderick <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 04/03/2003 11:55 AM Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: Log Files. I know this is not going to help BUT...there was a company in Europe and about 1 year ago I was having conversations with them about their MQ Log product. Amoung other things it would do exactly what you are requesting. Maybe someone from the company is eaves dropping on the LIST and will reply to you. Reviewing the product on a cursory level it seemed like the answer to some prayers!! bobbee >From: Ian Steyn <[EMAIL PROTECTED]> >Reply-To: MQSeries List <[EMAIL PROTECTED]> >To: [EMAIL PROTECTED] >Subject: Log Files. >Date: Thu, 3 Apr 2003 08:56:03 -0600 > >Hi Listers, > >We have a situation where some messages were "lost". >The applications says they did not receive the messages and the vendor >said they did send them. >I dumped the log files but it ended up in a 2.6G (gigabytes) file. > >My question: is there a way I can get the messages out of the log files >back onto a Q to read them back programmatically? >Or a tool that can handle a 2.6G log file. >This is NT MQ 5.2. > >Any help will be appreciated. > >Thanks >Ian _ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Disaster testing
One of the steps in the recovery procedure is to define a new BSDS data set and use REPRO to copy the most recent archived BSDS into it. My archive logs are on tape. Am I reading this correctly " If archive logs are on tape, The BSDS is the first data set of the first archive log volume. The BSDS is not repeated on later volmues." Does this mean I should use the CSQ.ARCHLOG1.A001 instead of CSQ.ARCHLOG1.B001? The instructions for archive logs on DASD are: If archive logs are on DASD, the BSDS is allocated on any available DASD. The BSDS name is like the corresponding archive log data set name; change only th e first letter of the last qualifier, from A to B, as in the example below: Archive log name CSQ.ARCHLOG1.A001 BSDS copy name CSQ.ARCHLOG1.B001 -Original Message- From: Robert Broderick [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 12:10 PM To: [EMAIL PROTECTED] Subject: Re: Disaster testing I think I pointed this out so if I am repeating myself I very much apollogize. I am Mainframe born and brought up. I now have my hand entrenched in distributed side processing. (six years of going back to school paid off!!). I would have to agree with Bump, I mean Lump, I mean HILL (tee hee hee). The distributed platforms TEND (and this is not a gerneralizationOH well...yes it is!!) not to think of DR in the same terms as the OS390 people do. Most of the times. I was in a shop that was doing financial transfers. If you designed a system that lost a transfer, you were looking for a good place to play checkers the next day. This was OS390. This also included a DR situation. In the past few contract with distributed platform they were happy to be back up and running!! What do you mean, I said in a meeting...you are running a brokerage system with DR rplication in ASYNC mode. BUT don't you understand. They said don't worry about it!! Drove me nuts!!! I couldn't sleep for a couple of nights. Well I digressed again. My origional point here is that to support true DR there will certainly be a good amoung of system (software side) to support the Holy Grail of DR physical recovery. The Designers must be aware of the limitations of the DR and must design redundency checking and transaction recovery into the system. DR is not perfect. But with both a physical and software approach it can become the Nervana you hope it to be. Please vote for me this coming Nov 11. bee-oh-dubble-bee-dubble-egh >From: "Hill, Dave" <[EMAIL PROTECTED]> >Reply-To: MQSeries List <[EMAIL PROTECTED]> >To: [EMAIL PROTECTED] >Subject: Re: Disaster testing >Date: Thu, 3 Apr 2003 11:27:46 -0500 > >As bee oh double beee double says: >It is a multi-platform muti-people environment. I have found that a service >level agreement ( SLA ) to be an option so that upon failure to recover you >can produce your SLA and show that it has been satisfied. Now that you have >seen the BS side of DR lets go on. >You need to recover all platforms and here in lays the problem. I have >recovered all platforms in testing but have yet to get all platform in >sync. This is a major goal but I feel it is the holly grail I seek. I am >"lucky" to be "in charge" of all MQ platforms and am the DRC too. I still >have trouble getting the admin of NT/AIX/WIN2000 machines to agree on a >single backup strategy. I still believe that MQSeries can be taken up a >notch in this area. >As for having a plan? I suggest test and test and fail and fail until >someone takes note. I find that non-Mainframe people do not understand what >true recovery is. You have to show them and hold their hands and cuddle and >whisper ( getting uneasy yet? ) until they see the light. >The bottom line is this. How do you define a point in time that can be >recovered and is acceptable to your business and or business units. >DR Its not just a job its an adventure ;-) > > >-Original Message- >From: mqm mqm [mailto:[EMAIL PROTECTED] >Sent: Thursday, April 03, 2003 10:01 AM >To: [EMAIL PROTECTED] >Subject: Re: Disaster testing > > >This is the first real discussion I've seen on this >list about DR. My favourite topic ...ha! ha! Everytime >you mention DR and MQ to IBM you get pointed to the >manual/redbook about restart/recovery. But as Bobbee >points out DR is potentially much more than this. >Which of Bobbee's three options are people adopting ? >Or are you doing something else ? Like just hoping you >never have a disaster ? > >mqm > >--- "Anderson, Lizette T. (RyTull)" ><[EMAIL PROTECTED]> wrote: > > Thanks for your response. There goes Vegas. I was > > thinking it may be > > easier to just start from scratch with new pagesets > > and logs. I am not even > > sure we can bring MQ down to get a decent backup. > > > > -Original Message- > > From: Robert Broderick > > [mailto:[EMAIL PROTECTED] > > Sent: Tuesday, April 01, 2003 2:09 PM > > To: [EMAIL PROTECTED] > > Subject: Re: Disaster testing > > > > >
Re: coredump when trying mqsicreatebroker
Thank you Derrick, the problem was solved with your recomendation. Regards, Yonny Serrano Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
The transaction will not trigger. One of the rules for first is that the queue is not opened for input. From: Ronald Weinger <[EMAIL PROTECTED]> Reply-To: MQSeries List <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: Re: Triggering Z/OS DLQH CSQUDLQH Date: Thu, 3 Apr 2003 15:54:29 -0500 If a CICS transaction, for example, starts due to trigger on first, and then does a GET WAIT UNLIMITED, it is possible than a new message on the queue will trigger a second execution of the transaction, but the current execution will retrieve the message before the second one is attached. So the second will get a 'no-message-available', and if coded to issue a GET WAIT UNLIMITED, will remain suspended. A third message will trigger another transaction, but it will be retrieved by one of the two currentl;y suspended transactions. That sequence could continue until max tasks are reached. However, have you found a MVS batch trigger monitor ? "Gary P. Klos" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Sent by: cc: "MQSeries List" Subject: Triggering Z/OS DLQH CSQUDLQH <[EMAIL PROTECTED] n.AC.AT> 04/03/2003 02:59 PM Please respond to "MQSeries List" I am setting up to use the Z/OS utility CSQUDLQH, which handles the dead letter queue for a queue manager. What I was thinking was to put a trigger on FIRST on the dead letter queue and once it starts CSQUDLQH, then I could leave it running indefinitely. My question is this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH removes all messages from the queue and stays running, then later another message goes on the dead letter queue, CSQUDLQH will start again. Does that mean I will have two batch jobs running CSQUDLQH, or does this program know to not run if it is already running with the same dead letter queue? Thanks, Gary Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive The information contained in this message may be CONFIDENTIAL and is for the intended addressee only. Any unauthorized use, dissemination of the information, or copying of this message is prohibited. If you are not the intended addressee, please notify the sender immediately and delete this message. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive _ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
Not true. In your example, IPPROCS would be 1. Another transaction will NOT be triggered inless no one has the queue open (IPPROCS = 0). My worry about triggering a DLH would be if the rules table said to retry a message. If the message's content was the problem, that message would probably wind up on the DLQ again. I could see loop potential there. If you're not retrying the messages, and just deleting them or maybe forwarding them to an application DLQ, that wouldn't be a problem. Ronald Weinger <[EMAIL PROTECTED]To: [EMAIL PROTECTED] .COM>cc: Sent by: MQSeriesSubject: Re: Triggering Z/OS DLQH CSQUDLQH List <[EMAIL PROTECTED] N.AC.AT> 04/03/2003 02:54 PM Please respond to MQSeries List If a CICS transaction, for example, starts due to trigger on first, and then does a GET WAIT UNLIMITED, it is possible than a new message on the queue will trigger a second execution of the transaction, but the current execution will retrieve the message before the second one is attached. So the second will get a 'no-message-available', and if coded to issue a GET WAIT UNLIMITED, will remain suspended. A third message will trigger another transaction, but it will be retrieved by one of the two currentl;y suspended transactions. That sequence could continue until max tasks are reached. However, have you found a MVS batch trigger monitor ? "Gary P. Klos" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Sent by: cc: "MQSeries List" Subject: Triggering Z/OS DLQH CSQUDLQH <[EMAIL PROTECTED] n.AC.AT> 04/03/2003 02:59 PM Please respond to "MQSeries List" I am setting up to use the Z/OS utility CSQUDLQH, which handles the dead letter queue for a queue manager. What I was thinking was to put a trigger on FIRST on the dead letter queue and once it starts CSQUDLQH, then I could leave it running indefinitely. My question is this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH removes all messages from the queue and stays running, then later another message goes on the dead letter queue, CSQUDLQH will start again. Does that mean I will have two batch jobs running CSQUDLQH, or does this program know to not run if it is already running with the same dead letter queue? Thanks, Gary Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive The information contained in this message may be CONFIDENTIAL and is for the intended addressee only. Any unauthorized use, dissemination of the information, or copying of this message is prohibited. If you are not the intended addressee, please notify the sender immediately and delete this message. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
If a CICS transaction, for example, starts due to trigger on first, and then does a GET WAIT UNLIMITED, it is possible than a new message on the queue will trigger a second execution of the transaction, but the current execution will retrieve the message before the second one is attached. So the second will get a 'no-message-available', and if coded to issue a GET WAIT UNLIMITED, will remain suspended. A third message will trigger another transaction, but it will be retrieved by one of the two currentl;y suspended transactions. That sequence could continue until max tasks are reached. However, have you found a MVS batch trigger monitor ? "Gary P. Klos" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Sent by: cc: "MQSeries List" Subject: Triggering Z/OS DLQH CSQUDLQH <[EMAIL PROTECTED] n.AC.AT> 04/03/2003 02:59 PM Please respond to "MQSeries List" I am setting up to use the Z/OS utility CSQUDLQH, which handles the dead letter queue for a queue manager. What I was thinking was to put a trigger on FIRST on the dead letter queue and once it starts CSQUDLQH, then I could leave it running indefinitely. My question is this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH removes all messages from the queue and stays running, then later another message goes on the dead letter queue, CSQUDLQH will start again. Does that mean I will have two batch jobs running CSQUDLQH, or does this program know to not run if it is already running with the same dead letter queue? Thanks, Gary Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive The information contained in this message may be CONFIDENTIAL and is for the intended addressee only. Any unauthorized use, dissemination of the information, or copying of this message is prohibited. If you are not the intended addressee, please notify the sender immediately and delete this message. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Triggering Z/OS DLQH CSQUDLQH
I had this same question regarding triggering. Apparently once a process has been triggered, the input use count is incremented and as long as it remains greater than zero, another process will not be triggered. Thanks, David Corbett IBM MQSeries Certified Specialist, Developer & Solutions Expert |-+---> | | "Gary P. Klos" | | | <[EMAIL PROTECTED]>| | | Sent by:| | | "MQSeries List" | | | <[EMAIL PROTECTED]| | | en.AC.AT> | | | | | | | | | 04/03/2003 01:59| | | PM | | | Please respond | | | to "MQSeries| | | List" | | | | |-+---> >---| | | |To: [EMAIL PROTECTED] | |cc: | |Subject: Triggering Z/OS DLQH CSQUDLQH | >---| I am setting up to use the Z/OS utility CSQUDLQH, which handles the dead letter queue for a queue manager. What I was thinking was to put a trigger on FIRST on the dead letter queue and once it starts CSQUDLQH, then I could leave it running indefinitely. My question is this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH removes all messages from the queue and stays running, then later another message goes on the dead letter queue, CSQUDLQH will start again. Does that mean I will have two batch jobs running CSQUDLQH, or does this program know to not run if it is already running with the same dead letter queue? Thanks, Gary Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: NetUserGetLocalGroups rc=1722 on NT4 SP3
Brian, Thanks for the feedback. We are looking into an upgrade to MQ for NT 5.3.1, but since that would have to go to production as well, it is obviously a laborious effort. Thanks for the help, I'll let everyone know what pans out. Dave Awerbuch > Date:Wed, 2 Apr 2003 11:56:36 -0500 > From:"Brian S. Crabtree" <[EMAIL PROTECTED]> > Subject: Re: NetUserGetLocalGroups rc=1722 on NT4 SP3 > > David > > I dont know what the answer to your particular problem is but > NetUserGetLocalGroups is the MS API route to allowing the domain mqm group > to have read access to everyone's group membership > > See > http://www.developer.ibm.com/tech/faq/results/0,1322,1%253A443%253A446%253A2 > 8%253Aobjgen,00.html > > The behaviour changed from NT4 to Win2000 > > See also > http://www.mqseries.net/phpBB/viewtopic.php?t=2383&highlight=1722 > > If you have MQSeries for Windows NT, V5.2 then I suggest you install the > latest CSD > > The issue is NT access security - it may be that the machine doesnt have > access to a domain controller, that the current logged on user doesnt have > mqm access, that MQ was installed from a user id with insufficient authority > etc > > Brian S. Crabtree > EAI Consultant > > - Original Message - > From: "David Awerbuch" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Wednesday, April 02, 2003 10:43 AM > Subject: NetUserGetLocalGroups rc=1722 on NT4 SP3 > > > > Hello all, > > > > Curious thing here. Every 10 seconds, I am getting this set of messages > in my > > amqerr01.log file. I just noticed t while trying to get Paul Clarke's > MO71 > > working in client mode against that machine, so I can;t say how long it's > been > > like this. > > > > -- > - > > 04/02/03 10:33:02 > > AMQ6119: An internal MQSeries error has occurred (rc=1722: > > NetUserGetLocalGroups: (null):) > > > > EXPLANATION: > > MQSeries detected an unexpected error when calling the operating system. > The > > MQSeries error recording routine has been called. > > ACTION: > > Use the standard facilities supplied with your system to record the > problem > > identifier, and to save the generated output files. Contact your IBM > support > > center. Do not discard these files until the problem has been resolved. > > -- > - > > 04/02/03 10:33:02 > > AMQ6184: An internal MQSeries error has occurred on queue manager > devprimeqm. > > > > EXPLANATION: > > An error has been detected, and the MQSeries error recording routine has > been > > called. The failing process is process 406. > > ACTION: > > Use the standard facilities supplied with your system to record the > problem > > identifier, and to save the generated output files. Contact your IBM > support > > center. Do not discard these files until the problem has been resolved. > > -- > - > > > > Process 406 is amqzlaa0.exe. I've already stopped and restarted the queue > > manager, but with no change. > > > > Does anyone have an idea? > > > > Please be sure to reply directly to me > mailto:[EMAIL PROTECTED] > > as well as the newsgroup; I only get digests. > > > > > > Thanks, > > Dave A. > > [EMAIL PROTECTED] > > __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Triggering Z/OS DLQH CSQUDLQH
I am setting up to use the Z/OS utility CSQUDLQH, which handles the dead letter queue for a queue manager. What I was thinking was to put a trigger on FIRST on the dead letter queue and once it starts CSQUDLQH, then I could leave it running indefinitely. My question is this, if I use TRIGGER ON FIRST, and the queue triggers then CSQUDLQH removes all messages from the queue and stays running, then later another message goes on the dead letter queue, CSQUDLQH will start again. Does that mean I will have two batch jobs running CSQUDLQH, or does this program know to not run if it is already running with the same dead letter queue? Thanks, Gary Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Shared and non-shared ques on a single que manager?
Yes you definitely can. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Shared and non-shared ques on a single que manager?
Absolutely, you can mix shared and non-shared queues on a single qmgr in an OS/390 or z/OS sysplex environment. I think 63k messages are fairly large and you should consider breaking them into smaller pieces if possible. You can easily flood the network if you are going to be sending a large number of these messages. Thanks, David Corbett Work: 651-205-2714 Pager: 651-610-3842 |-+---> | | "Perfetto, Ed" | | | <[EMAIL PROTECTED]| | | LINC.COM> | | | Sent by:| | | "MQSeries List" | | | <[EMAIL PROTECTED]| | | en.AC.AT> | | | | | | | | | 04/03/2003 01:26| | | PM | | | Please respond | | | to "MQSeries| | | List" | | | | |-+---> >---| | | |To: [EMAIL PROTECTED] | |cc: | |Subject: Shared and non-shared ques on a single que manager? | >---| New subscriber here!! Can you or can you not have shared and non-shared ques on a single que manager? We're trying to overcome the 63K limits in a sysplex environment and have decided to have multiple que managers, some shared(<63K) and others not (>63K). Thanks, Ed. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
Qpasa monitors this as well. -Original Message- From: Diwakar S Yammanuru [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 12:21 PM To: [EMAIL PROTECTED] Subject: Re: How to Monitor queue size full Candle does monitor pageset usage on OS/390. I am not sure about Unix. From: "mqm mqm" <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 08:30 AM PST Please respond to "MQSeries List" <[EMAIL PROTECTED]> Sent by: "MQSeries List" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size full It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Shared and non-shared ques on a single que manager?
New subscriber here!! Can you or can you not have shared and non-shared ques on a single que manager? We're trying to overcome the 63K limits in a sysplex environment and have decided to have multiple que managers, some shared(<63K) and others not(>63K). Thanks, Ed.
Re: How to Monitor queue size full
Candle does offer, for an additional cost of course, another agent that does monitor cpu, filesystem space, etc. Diwakar S Yammanuru cc: Sent by: MQSeries Subject: Re: How to Monitor queue size full List <[EMAIL PROTECTED] AC.AT> 04/03/2003 01:21 PM Please respond to MQSeries List Candle does monitor pageset usage on OS/390. I am not sure about Unix. From: "mqm mqm" <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 08:30 AM PST Please respond to "MQSeries List" <[EMAIL PROTECTED]> Sent by: "MQSeries List" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size full It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Authorizations in 5.3
George, I am not sure that I follow this example. Applications should only need authorization on the QREMOTE or QALIAS queues. There is no need to authorize applications to the transmission queue. -Original Message-From: George Beaton [mailto:[EMAIL PROTECTED]Sent: Wednesday, April 02, 2003 5:57 PMTo: [EMAIL PROTECTED]Subject: Re: Authorizations in 5.3There's another little problem with it that I overlooked. If you connect to a queue manager other than the one over which the security profiles are implemented then you can issue an MQOPEN using ObjectName ObjectQmanagerName and follow up with a series of MQPUTs to the queue. Consequently you need to put security onto every transmission queue to the secured Queue managerGeorge BeatonDB&MW62448125 spectrum 148125 "Jones, Ralph C {PBSG}" <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 03/04/2003 08:37 Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: Authorizations in 5.3 The REFRESH SECURITY command was introduced with 5.2 on the distributed codebase.The behavior concerning authorization being granted to all members of asecurity group when issued to a member only occurs in UNIX environments.-Original Message-From: Neil Casey [mailto:[EMAIL PROTECTED]Sent: Tuesday, April 01, 2003 5:54 PMTo: [EMAIL PROTECTED]Subject: Re: Authorizations in 5.3Hi Dave,REFRESH SECURITY was not available to unix or windows systems until 5.3. Ithas always been implemented on zOS and was used to refresh the RACFprofiles in storage.The security capabilities of 5.3 have been extended quite a bit. So much sothat there is now a book dedicated to just that topic.You might want to review it, especially looking at generic queue names insecurity profiles, as well as the REFRESH requirements.Another thing that can be confusing about the security cache is that itdoes not get preloaded. I believe that security details for an object areonly loaded when the object is referenced by the particular user. Thismeans that if you define the security before the object is referenced, itwill work without a refresh, but once the user has attempted to access theobject and the access failed, you need to run REFRESH, or restart the queuemanager, to get it to work.You can look at all your current profiles with amqoamd or dmpmqaut (theyreturn the same info in different formats).Also, be careful of defining new objects using a user other than mqm. Ifthe primary group of user 'neil' is 'staff', and user 'neil' is a member of'mqm', then 'neil' can use runmqsc and define objects. However, the 'staff'group is automatically given +all access to those objects. This is not abug because it is documented to work this way in the manuals. It just isn'tvery sensible in my opinion. Is it possible that this behaviour is causingthe confusion in results that you are seeing?Rgeards,Neil Casey.|-+>| | "De Seve, David" || | <[EMAIL PROTECTED]|| | RIOTT.COM> || | Sent by: MQSeries|| | List || | <[EMAIL PROTECTED]|| | n.AC.AT> || | || | || | 02/04/2003 02:31 || | Please respond to|| | MQSeries List || | ||-+>>--| || | To: [EMAIL PROTECTED]| | cc:| | Subject: Re: Authorizations in 5.3|>--|By the way we created a new queue today and again tried to access it(without running the command below) and we had no issue? This appears toberandom behavior.-Original Message-From: De Seve, David [mailto:[EMAIL PROTECTED]Sent: Tuesday, April 01, 2003 11:26 AM To: [EMAIL PROTECTED]Subject: Re: Authorizations in 5.3That's not a silly question. Is that a new command for 5.3?-Original Message-From: Neil Casey [mailto:[EMAIL PROTECTED]Sent: Monday, March 31, 2003 5:54 PMTo: [EMAIL PROTECTED]Subject: Re: Authorizations in 5.3Hi Dave, Silly question, but did you run REFRESH SECURITY(*) in runmqsc orrestart your queue manager?Neil Casey|-+>| | "De Seve, David" || | <[EMAIL PROTECTED]|| | RIOTT.COM> || | Sent by: MQSeries|| | List
Re: XML over HTTP and MQ
SOAP does not describe the transport method. SOAP can be implemented over HTTP, SMTP, etc. However, these protocols lack the "assuredness" WMQ provides. I believe future releases of HTTP will include assured delivery and other QOS features. >>> [EMAIL PROTECTED] 04/03/03 11:07AM >>> I think a lot of people consider Web Services to be a standard that provides for 1) a means of describing how a 'service' will be called (a 'contract' that describes a function(s), the response(s) and the data types of the parameters), and 2) a means of providing a place (directory) where these service descriptions can be accessed. If you think of SOAP as the set of XML that describes the function call, as well as the actual content/payload, then you can relate object oriented programming to SOAP and Web Services by saying that Web Services is the class, and a SOAP message is the instance of the class. I know that's not quite right, but the point is that SOAP is the message content. It could go over MQ or over HTTP. You still have to send and receive it. Right now SOAP listeners are built into HTTP servers in most cases. I'm not familiar with the MQ SOAP listener.. Someone help me out here.. Does the MQ SOAP listener actually examine the SOAP content, determine the application and start it, passing the SOAP Body to it?? I vaguely remember that you have to 'predefine' the application for the SOAP listener per queue.. Something has to translate the XML-described request for a function call, into the native environment's "call a program" process. If not, then there is still customization to be done. SOAP over HTTP is probably easier just because it is 'built in' to a lot of web browsers/app servers. I could be wrong but SOAP over MQ is just as standardized as SOAP over HTTP; it's what happens at the endpoint that is proprietary/customized. -Original Message- From: mqm mqm [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 10:06 AM To: [EMAIL PROTECTED] Subject: Re: XML over HTTP and MQ If you do SOAP over MQ does that mean you're not doing Web Services because you are using a proprietary protocol (MQ) rather than an open internet standard protocol (http). Views anyone ? mqm --- "Heggie, Peter" <[EMAIL PROTECTED]> wrote: > It lacks a bunch of things.. > > XML over HTTP - when you don't care enough to send > the very best.. > > -Original Message- > From: Jason Cornell > [mailto:[EMAIL PROTECTED] > Sent: Wednesday, April 02, 2003 1:13 PM > To: [EMAIL PROTECTED] > Subject: Re: XML over HTTP and MQ > > > HTTP lacks WMQ's "assured" delivery. > > >>> [EMAIL PROTECTED] 04/02/03 08:29AM >>> > Where does or how does MQ fit into the XML over HTTP > space ? Seems like this > method potentially does not need MQ as the > middleware. > > > > Wesley Shaw > OJRP 10th Floor > Work: 804 771 3589 (736-3589) > Pager: 888 436 2805 Mobile 804 512 5260 > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > This e-mail and any files transmitted with it, are > confidential to National Grid and are intended > solely for the use of the individual or entity to > whom they are addressed. If you have received this > e-mail in error, please contact the National Grid > USA Enterprise Support Center on 508-389-3375 or > 315-428-6360. > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
Candle does monitor pageset usage on OS/390. I am not sure about Unix. From: "mqm mqm" <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 08:30 AM PST Please respond to "MQSeries List" <[EMAIL PROTECTED]> Sent by: "MQSeries List" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size full It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Disaster testing
I think I pointed this out so if I am repeating myself I very much apollogize. I am Mainframe born and brought up. I now have my hand entrenched in distributed side processing. (six years of going back to school paid off!!). I would have to agree with Bump, I mean Lump, I mean HILL (tee hee hee). The distributed platforms TEND (and this is not a gerneralizationOH well...yes it is!!) not to think of DR in the same terms as the OS390 people do. Most of the times. I was in a shop that was doing financial transfers. If you designed a system that lost a transfer, you were looking for a good place to play checkers the next day. This was OS390. This also included a DR situation. In the past few contract with distributed platform they were happy to be back up and running!! What do you mean, I said in a meeting...you are running a brokerage system with DR rplication in ASYNC mode. BUT don't you understand. They said don't worry about it!! Drove me nuts!!! I couldn't sleep for a couple of nights. Well I digressed again. My origional point here is that to support true DR there will certainly be a good amoung of system (software side) to support the Holy Grail of DR physical recovery. The Designers must be aware of the limitations of the DR and must design redundency checking and transaction recovery into the system. DR is not perfect. But with both a physical and software approach it can become the Nervana you hope it to be. Please vote for me this coming Nov 11. bee-oh-dubble-bee-dubble-egh From: "Hill, Dave" <[EMAIL PROTECTED]> Reply-To: MQSeries List <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: Re: Disaster testing Date: Thu, 3 Apr 2003 11:27:46 -0500 As bee oh double beee double says: It is a multi-platform muti-people environment. I have found that a service level agreement ( SLA ) to be an option so that upon failure to recover you can produce your SLA and show that it has been satisfied. Now that you have seen the BS side of DR lets go on. You need to recover all platforms and here in lays the problem. I have recovered all platforms in testing but have yet to get all platform in sync. This is a major goal but I feel it is the holly grail I seek. I am "lucky" to be "in charge" of all MQ platforms and am the DRC too. I still have trouble getting the admin of NT/AIX/WIN2000 machines to agree on a single backup strategy. I still believe that MQSeries can be taken up a notch in this area. As for having a plan? I suggest test and test and fail and fail until someone takes note. I find that non-Mainframe people do not understand what true recovery is. You have to show them and hold their hands and cuddle and whisper ( getting uneasy yet? ) until they see the light. The bottom line is this. How do you define a point in time that can be recovered and is acceptable to your business and or business units. DR Its not just a job its an adventure ;-) -Original Message- From: mqm mqm [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 10:01 AM To: [EMAIL PROTECTED] Subject: Re: Disaster testing This is the first real discussion I've seen on this list about DR. My favourite topic ...ha! ha! Everytime you mention DR and MQ to IBM you get pointed to the manual/redbook about restart/recovery. But as Bobbee points out DR is potentially much more than this. Which of Bobbee's three options are people adopting ? Or are you doing something else ? Like just hoping you never have a disaster ? mqm --- "Anderson, Lizette T. (RyTull)" <[EMAIL PROTECTED]> wrote: > Thanks for your response. There goes Vegas. I was > thinking it may be > easier to just start from scratch with new pagesets > and logs. I am not even > sure we can bring MQ down to get a decent backup. > > -Original Message- > From: Robert Broderick > [mailto:[EMAIL PROTECTED] > Sent: Tuesday, April 01, 2003 2:09 PM > To: [EMAIL PROTECTED] > Subject: Re: Disaster testing > > > I am not an expert on this but here is my > 2.314567489364532 cents. > > Well there are different levels/approaches to what > people call DR. Some > believe it is an image restore of the queue manager > at the point of failure. > Others see it as COB (Continuance of Business). The > two views are radically > different from a hardware (and yes! Software) stand > point. > > If you are viewing it as recoverable from a point in > time then all > participants must agree on a point in time. Usually > this is a climb up the > staircase of the tower of babble. If you can get all > the platform people to > sit in a room and agree on a coordination to achieve > this. I will take you > to Vegas for the weekend!!! As you see, I'm not > worried about loosing my > shirt at the tables because to achieve this is > almost impossible. > > The lowest point of view is the COB one. Do you want > to put in place a > framework where you can, with a bit of effort, bring > the business back up to > operating speed to move forward and your business > a
Re: Log Files.
I know this is not going to help BUT...there was a company in Europe and about 1 year ago I was having conversations with them about their MQ Log product. Amoung other things it would do exactly what you are requesting. Maybe someone from the company is eaves dropping on the LIST and will reply to you. Reviewing the product on a cursory level it seemed like the answer to some prayers!! bobbee From: Ian Steyn <[EMAIL PROTECTED]> Reply-To: MQSeries List <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: Log Files. Date: Thu, 3 Apr 2003 08:56:03 -0600 Hi Listers, We have a situation where some messages were "lost". The applications says they did not receive the messages and the vendor said they did send them. I dumped the log files but it ended up in a 2.6G (gigabytes) file. My question: is there a way I can get the messages out of the log files back onto a Q to read them back programmatically? Or a tool that can handle a 2.6G log file. This is NT MQ 5.2. Any help will be appreciated. Thanks Ian _ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
There is NO reason to monitor queue sizes. The queue files (e.g., AIX platform is /var/mqm/qmgrs//queues/*) DYNAMICALLY change in size based on usage or when a queue manager is restarted or if previously active queues have been idle for some time. What you mean to say is "you want to monitor the size of the file system where the queues for the queue manager in question reside. I believe under V5.1 and below, queue sizes could grow to 320MBs, V5.2 this changed to 2GBs and V5.3 its now 2TBs (for UNIX systems). For V5.1 systems, in the TuningParameters section of the mq.ini file, if you specify the keyword 'DefaultQFileSize' and set it to 1GB (DefaultQFileSize=10), then bounce the queue manager, any NEW queues created will now grow up to this size. If a "Queue Full" condition is encountered, check messages that go to your DLQ for a reason code of 2053 (MQRC_Q_FULL or Hexadecimal number 0805 - how the RC is stored in the DLQ header). Now tailor your DLQ Rules Table appropriately for the action you want your DLQ utility (e.g., runmqdlq) to take when this type of error condition is encountered. REASON(MQRC_Q_FULL) ACTION(FWD) FEWQ(OverFlowQueue) runmqdlq < MyRulesTable.txt A 2053 reason code is normally associated with "maximum number of messages reached". I have never seen a condition where the queue size was at capacity. If somebody has seen this condition and it was registered within the AMQERR01.LOG file, please respond back as to what the AMQ number issue. I'm sure I would not be the only one interested in this error. We do manage messages UP TO 85MBs in size, so if the application reading this queue becomes hung, it won't take but a few hundred messages before this condition becomes reality. Also check your SYSTEM.ADMIN.PERFM.EVENT queue, because when queue full conditions are encountered, an event message will be put to this queue by the queue manager that detected this condition. If you have enabled PERFORMANCE EVENTS and are monitoring this queue, then you should be able to catch this condition when it occurs, and alert your staff. Stephan C. Moen -Original Message- From: MQSeries List [mailto:[EMAIL PROTECTED] Behalf Of Christopher D. Fryett Sent: Thursday, April 03, 2003 10:05 AM To: [EMAIL PROTECTED] Subject: Re: How to Monitor queue size full If I recall correctly there is no event triggers for queue size, just depth. If you are coming across an issue of queue size you may want to look at your design again and validate you can support the message sizes for the queue definitions you have. You may also need to review the method in how you put and get information on the queues. Unless there is a specific reason to leave the messages on the queues, you may want to figure out (if it is an issue) why they are not being pulled from the queue ASAP. Chris -Original Message- From: MQSeries List [mailto:[EMAIL PROTECTED] On Behalf Of Jiede J Yang Sent: Thursday, April 03, 2003 9:06 AM To: [EMAIL PROTECTED] Subject: Re: How to Monitor queue size full Look into the Event Monitor book in the MQ Info Center, you can rely on soultino from there Thanks. Email: [EMAIL PROTECTED] mqm mqm <[EMAIL PROTECTED]To: [EMAIL PROTECTED] O.UK>cc: Sent by: MQSeriesSubject: Re: How to Monitor queue size full List <[EMAIL PROTECTED] N.AC.AT> 04/03/2003 09:48 AM Please respond to MQSeries List There's a number of tools that monitor system resource usage or alert when thresholds are reached e.g. BMC Patrol, Tivoli. Any particular platform ? mqm --- "Mittal, Gaurav" <[EMAIL PROTECTED]> wrote: > We have queues defined to have a maximum depth of > 6,40,000 messages and a > max size of 2GB.I know we can monitor the queue > depths using 'High Depth > Event' .Is there a way by which we can monitor the > queue size (in terms of > file size not number of messages) also. > Any pointers would be useful.. > > Thanks n Regds > Gaurav > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your
Re: How to Monitor queue size full
So does BMC's product. Peter Armaly - Senior Software Consultant bmcsoftware, Toronto Work: 905-602-2839 Mobile: 416-998-6748 email: [EMAIL PROTECTED] -Original Message-From: Glen Shubert [mailto:[EMAIL PROTECTED]Sent: Thursday, April 03, 2003 11:58 AMTo: [EMAIL PROTECTED]Subject: Re: How to Monitor queue size fullThe Candle product does monitor Page Dataset usage as to percentage full, etcGlen ShubertAssociate DirectorTSYS - MQSeries Technical SupportIBM Certified Specialist - MQSeriesIBM Certified Solutions Expert - MQSeries mqm mqm <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 04/03/2003 11:30 AM Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size fullIt's not the MQ objects Gaurav needs to monitor. It'sthe underlying file usage. So on Unix how much spaceis available in the /var/mqm directory for example. OnOS/390 it would be the pageset usage. Correct me ifI'm wrong but I don't believe Candle or any of theother MQ specific monitors (e.g. QPasa, Nastel) dothis.mqm--- Diwakar S Yammanuru <[EMAIL PROTECTED]>wrote:> The product Candle can monitor MQ objects and issue> alerts on many> platforms.You may visit www.candle.com>> Thanks.> Diwakar. From: "mqm mqm"> <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003> 06:48 AM> PST>> Please respond to "MQSeries List"> <[EMAIL PROTECTED]>>> Sent by: "MQSeries List" <[EMAIL PROTECTED] To: [EMAIL PROTECTED]> cc:> Subject: Re: How to Monitor queue size full>>> There's a number of tools that monitor system> resource> usage or alert when thresholds are reached e.g. BMC> Patrol, Tivoli. Any particular platform ?>> mqm>> --- "Mittal, Gaurav" <[EMAIL PROTECTED]>> wrote:> > We have queues defined to have a maximum depth of> > 6,40,000 messages and a> > max size of 2GB.I know we can monitor the queue> > depths using 'High Depth> > Event' .Is there a way by which we can monitor the> > queue size (in terms of> > file size not number of messages) also.> > Any pointers would be useful..> >> > Thanks n Regds> > Gaurav> >> > Instructions for managing your mailing list> > subscription are provided in> > the Listserv General Users Guide available at> > http://www.lsoft.com> > Archive: http://vm.akh-wien.ac.at/MQSeries.archive>>> __> Do you Yahoo!?> Yahoo! Tax Center - File online, calculators, forms,> and more> http://tax.yahoo.com>> Instructions for managing your mailing list> subscription are provided in> the Listserv General Users Guide available at> http://www.lsoft.com> Archive: http://vm.akh-wien.ac.at/MQSeries.archive>> Instructions for managing your mailing list> subscription are provided in> the Listserv General Users Guide available at> http://www.lsoft.com> Archive: http://vm.akh-wien.ac.at/MQSeries.archive__Do you Yahoo!?Yahoo! Tax Center - File online, calculators, forms, and morehttp://tax.yahoo.comInstructions for managing your mailing list subscription are provided inthe Listserv General Users Guide available at http://www.lsoft.comArchive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
The Candle product does monitor Page Dataset usage as to percentage full, etc Glen Shubert Associate Director TSYS - MQSeries Technical Support IBM Certified Specialist - MQSeries IBM Certified Solutions Expert - MQSeries mqm mqm <[EMAIL PROTECTED]> Sent by: MQSeries List <[EMAIL PROTECTED]> 04/03/2003 11:30 AM Please respond to MQSeries List To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size full It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
Just to cover one base. At clients, I have scripts that I install that monitor EVERYTHIN from file space to processes to queues to channels to MQSI to adapters. Failures get EMAILs and pages. Works great and runs off CRON. bobbee From: mqm mqm <[EMAIL PROTECTED]> Reply-To: MQSeries List <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] Subject: Re: How to Monitor queue size full Date: Thu, 3 Apr 2003 08:30:29 -0800 It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive _ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=features/junkmail Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
It's not the MQ objects Gaurav needs to monitor. It's the underlying file usage. So on Unix how much space is available in the /var/mqm directory for example. On OS/390 it would be the pageset usage. Correct me if I'm wrong but I don't believe Candle or any of the other MQ specific monitors (e.g. QPasa, Nastel) do this. mqm --- Diwakar S Yammanuru <[EMAIL PROTECTED]> wrote: > The product Candle can monitor MQ objects and issue > alerts on many > platforms.You may visit www.candle.com > > Thanks. > Diwakar. > > > > From: "mqm mqm" > <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 > 06:48 AM > PST > > Please respond to "MQSeries List" > <[EMAIL PROTECTED]> > > Sent by: "MQSeries List" <[EMAIL PROTECTED]> > > > To: [EMAIL PROTECTED] > cc: > Subject: Re: How to Monitor queue size full > > > There's a number of tools that monitor system > resource > usage or alert when thresholds are reached e.g. BMC > Patrol, Tivoli. Any particular platform ? > > mqm > > --- "Mittal, Gaurav" <[EMAIL PROTECTED]> > wrote: > > We have queues defined to have a maximum depth of > > 6,40,000 messages and a > > max size of 2GB.I know we can monitor the queue > > depths using 'High Depth > > Event' .Is there a way by which we can monitor the > > queue size (in terms of > > file size not number of messages) also. > > Any pointers would be useful.. > > > > Thanks n Regds > > Gaurav > > > > Instructions for managing your mailing list > > subscription are provided in > > the Listserv General Users Guide available at > > http://www.lsoft.com > > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > __ > Do you Yahoo!? > Yahoo! Tax Center - File online, calculators, forms, > and more > http://tax.yahoo.com > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Disaster testing
As bee oh double beee double says: It is a multi-platform muti-people environment. I have found that a service level agreement ( SLA ) to be an option so that upon failure to recover you can produce your SLA and show that it has been satisfied. Now that you have seen the BS side of DR lets go on. You need to recover all platforms and here in lays the problem. I have recovered all platforms in testing but have yet to get all platform in sync. This is a major goal but I feel it is the holly grail I seek. I am "lucky" to be "in charge" of all MQ platforms and am the DRC too. I still have trouble getting the admin of NT/AIX/WIN2000 machines to agree on a single backup strategy. I still believe that MQSeries can be taken up a notch in this area. As for having a plan? I suggest test and test and fail and fail until someone takes note. I find that non-Mainframe people do not understand what true recovery is. You have to show them and hold their hands and cuddle and whisper ( getting uneasy yet? ) until they see the light. The bottom line is this. How do you define a point in time that can be recovered and is acceptable to your business and or business units. DR Its not just a job its an adventure ;-) -Original Message- From: mqm mqm [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 10:01 AM To: [EMAIL PROTECTED] Subject: Re: Disaster testing This is the first real discussion I've seen on this list about DR. My favourite topic ...ha! ha! Everytime you mention DR and MQ to IBM you get pointed to the manual/redbook about restart/recovery. But as Bobbee points out DR is potentially much more than this. Which of Bobbee's three options are people adopting ? Or are you doing something else ? Like just hoping you never have a disaster ? mqm --- "Anderson, Lizette T. (RyTull)" <[EMAIL PROTECTED]> wrote: > Thanks for your response. There goes Vegas. I was > thinking it may be > easier to just start from scratch with new pagesets > and logs. I am not even > sure we can bring MQ down to get a decent backup. > > -Original Message- > From: Robert Broderick > [mailto:[EMAIL PROTECTED] > Sent: Tuesday, April 01, 2003 2:09 PM > To: [EMAIL PROTECTED] > Subject: Re: Disaster testing > > > I am not an expert on this but here is my > 2.314567489364532 cents. > > Well there are different levels/approaches to what > people call DR. Some > believe it is an image restore of the queue manager > at the point of failure. > Others see it as COB (Continuance of Business). The > two views are radically > different from a hardware (and yes! Software) stand > point. > > If you are viewing it as recoverable from a point in > time then all > participants must agree on a point in time. Usually > this is a climb up the > staircase of the tower of babble. If you can get all > the platform people to > sit in a room and agree on a coordination to achieve > this. I will take you > to Vegas for the weekend!!! As you see, I'm not > worried about loosing my > shirt at the tables because to achieve this is > almost impossible. > > The lowest point of view is the COB one. Do you want > to put in place a > framework where you can, with a bit of effort, bring > the business back up to > operating speed to move forward and your business > and IT people will hammer > out what was going on at the time of the disaster > and fix it with their > magic wands, special programs to sync up the servers > and catch isolated > transaction from replaying and lots of bubblegum. > This is usually the > easiest and looks to be what you are talking about. > Remember in a situation > like this the servers aren't in sync so there may be > duplicates. If it's a > financial system you are going to need LOTS of > bubblegum!! > > A third possibility that I failed to mention is a > hardware solution. This > being where you have your OS390 fiber connected to a > redundant site, your > UNIX servers Veritas clustered AND replicated in > SYNC and the NT boxes with > a similar redundancy. I'm not talking about just > mirrored disk arrays here. > IF you run a DR then the boxes failover to their > mirrored replicated > partners and the day moves ahead with minimal > outage. This is EXPENSIVE from > a hardware and support staff point of view. (Someone > has to know how to set > up and monitor this stuff). > > AS I said prior, Hardware, Software or platform > coordination. Pick your > poison, it is all downhill from here with a big > bottle of aspirins waiting > for you at the bottom. > > bobbee > > > > > > > >From: "Anderson, Lizette T. (RyTull)" > <[EMAIL PROTECTED]> > >Reply-To: MQSeries List <[EMAIL PROTECTED]> > >To: [EMAIL PROTECTED] > >Subject: Disaster testing > >Date: Tue, 1 Apr 2003 13:15:03 -0600 > > > >We are in the process of creating a disaster plan > for MQSeries v 5.2 > >running > >on OS/390, Windows 2000 and AS/400. I wanted to > get some ideas. The plan > >is to run a backup of MQS
Re: XML over HTTP and MQ
I think a lot of people consider Web Services to be a standard that provides for 1) a means of describing how a 'service' will be called (a 'contract' that describes a function(s), the response(s) and the data types of the parameters), and 2) a means of providing a place (directory) where these service descriptions can be accessed. If you think of SOAP as the set of XML that describes the function call, as well as the actual content/payload, then you can relate object oriented programming to SOAP and Web Services by saying that Web Services is the class, and a SOAP message is the instance of the class. I know that's not quite right, but the point is that SOAP is the message content. It could go over MQ or over HTTP. You still have to send and receive it. Right now SOAP listeners are built into HTTP servers in most cases. I'm not familiar with the MQ SOAP listener.. Someone help me out here.. Does the MQ SOAP listener actually examine the SOAP content, determine the application and start it, passing the SOAP Body to it?? I vaguely remember that you have to 'predefine' the application for the SOAP listener per queue.. Something has to translate the XML-described request for a function call, into the native environment's "call a program" process. If not, then there is still customization to be done. SOAP over HTTP is probably easier just because it is 'built in' to a lot of web browsers/app servers. I could be wrong but SOAP over MQ is just as standardized as SOAP over HTTP; it's what happens at the endpoint that is proprietary/customized. -Original Message- From: mqm mqm [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 10:06 AM To: [EMAIL PROTECTED] Subject: Re: XML over HTTP and MQ If you do SOAP over MQ does that mean you're not doing Web Services because you are using a proprietary protocol (MQ) rather than an open internet standard protocol (http). Views anyone ? mqm --- "Heggie, Peter" <[EMAIL PROTECTED]> wrote: > It lacks a bunch of things.. > > XML over HTTP - when you don't care enough to send > the very best.. > > -Original Message- > From: Jason Cornell > [mailto:[EMAIL PROTECTED] > Sent: Wednesday, April 02, 2003 1:13 PM > To: [EMAIL PROTECTED] > Subject: Re: XML over HTTP and MQ > > > HTTP lacks WMQ's "assured" delivery. > > >>> [EMAIL PROTECTED] 04/02/03 08:29AM >>> > Where does or how does MQ fit into the XML over HTTP > space ? Seems like this > method potentially does not need MQ as the > middleware. > > > > Wesley Shaw > OJRP 10th Floor > Work: 804 771 3589 (736-3589) > Pager: 888 436 2805 Mobile 804 512 5260 > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > This e-mail and any files transmitted with it, are > confidential to National Grid and are intended > solely for the use of the individual or entity to > whom they are addressed. If you have received this > e-mail in error, please contact the National Grid > USA Enterprise Support Center on 508-389-3375 or > 315-428-6360. > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Max Channels
Jerry, I would look at the log files and see if there are any error messages which might help. Also, there is a TCP keep alive setting for the qmgr in the Services menu. This needs to be set for the qmgr to use keep alive. I'm not sure keep alive will do anything more than clean up connections which have not been properly disconnected by the application. The max channel settings do not ensure that you will get 1000 connections. This will depend on the resources of the server (memory and CPU). There is also another setting in the TCP panel of the services menu ListenerBacklog (the number of backlogged TCP requests for MQ listener service). The default setting for windows is 100. A 9228 error message will appear on the log if this setting is exceeded. Nick -Original Message- From: Jiede J Yang [mailto:[EMAIL PROTECTED] Sent: Thursday, April 03, 2003 7:41 AM To: [EMAIL PROTECTED] Subject: Max Channels Importance: High All: Background: MQ (5.3.0.1) Server is running on a Win2k Server machine. There is one MQ Q manager for this issue. There are 3 local Qs in the Q manager. The Max channel and max active channels has been set to 1000. A local Win2k Service was used to put message into the qs. Client is Java also on Win2k Server. There are two physical machines with exactly the same Java code (Websphere Application Server). When the app server started, 6 threads are created. 1. Each thread will obtained a Q Manager 2. Loop For Ever (Get Msgs) 2.1: Access a Q from the Q manager 2.2: get msg with the Q 2.3 close the Q. 3. When the Application Server shuts down, the Q manager will be disconnected. Note: If there is a problem during the loop, a reconnect logic will be done to the Q manager. The reconnection is a loop by itself with a interval of 3 sec until I get a good connection. Problem: Due to whatever reason, a reconnection logic is triggered, then we will run out of connection (even I have configured the number to be 1000). Trail Fix: Check the TCP tab's KeepAlive Check Box in the WebSphere MQ Service Snap-in. Change the KeepLiveTime in Registry. (HKEY_ LOCAL_ MACHINE\ SYSTEM\ CurrentControlSet\Services\ Tcpip\ Parameters; KeepAliveTime set to 6). We tested this fix, it seems not working as expected. Any other fix ideas? Thanks. Jerry [EMAIL PROTECTED] Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
If I recall correctly there is no event triggers for queue size, just depth. If you are coming across an issue of queue size you may want to look at your design again and validate you can support the message sizes for the queue definitions you have. You may also need to review the method in how you put and get information on the queues. Unless there is a specific reason to leave the messages on the queues, you may want to figure out (if it is an issue) why they are not being pulled from the queue ASAP. Chris -Original Message- From: MQSeries List [mailto:[EMAIL PROTECTED] On Behalf Of Jiede J Yang Sent: Thursday, April 03, 2003 9:06 AM To: [EMAIL PROTECTED] Subject: Re: How to Monitor queue size full Look into the Event Monitor book in the MQ Info Center, you can rely on soultino from there Thanks. Email: [EMAIL PROTECTED] mqm mqm <[EMAIL PROTECTED]To: [EMAIL PROTECTED] O.UK>cc: Sent by: MQSeriesSubject: Re: How to Monitor queue size full List <[EMAIL PROTECTED] N.AC.AT> 04/03/2003 09:48 AM Please respond to MQSeries List There's a number of tools that monitor system resource usage or alert when thresholds are reached e.g. BMC Patrol, Tivoli. Any particular platform ? mqm --- "Mittal, Gaurav" <[EMAIL PROTECTED]> wrote: > We have queues defined to have a maximum depth of > 6,40,000 messages and a > max size of 2GB.I know we can monitor the queue > depths using 'High Depth > Event' .Is there a way by which we can monitor the > queue size (in terms of > file size not number of messages) also. > Any pointers would be useful.. > > Thanks n Regds > Gaurav > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Disaster testing
I think it all depends on what you agree upon with your users, e.g. SLA. Which, in turns, influences the approach you take for DR. For those who require real-time redundancy, they spend the bucks for replicating data to remote sites. For others, it might be sufficient to restore to the last business day. In either case, there's no gurantee for 100% data integrity, but the big spender will be a lot closer to it. mqm mqm <[EMAIL PROTECTED] To: [EMAIL PROTECTED] CO.UK> cc: Sent by: Subject: Re: Disaster testing MQSeries List <[EMAIL PROTECTED] en.AC.AT> 04/03/2003 10:00 AM Please respond to MQSeries List This is the first real discussion I've seen on this list about DR. My favourite topic ...ha! ha! Everytime you mention DR and MQ to IBM you get pointed to the manual/redbook about restart/recovery. But as Bobbee points out DR is potentially much more than this. Which of Bobbee's three options are people adopting ? Or are you doing something else ? Like just hoping you never have a disaster ? mqm --- "Anderson, Lizette T. (RyTull)" <[EMAIL PROTECTED]> wrote: > Thanks for your response. There goes Vegas. I was > thinking it may be > easier to just start from scratch with new pagesets > and logs. I am not even > sure we can bring MQ down to get a decent backup. > > -Original Message- > From: Robert Broderick > [mailto:[EMAIL PROTECTED] > Sent: Tuesday, April 01, 2003 2:09 PM > To: [EMAIL PROTECTED] > Subject: Re: Disaster testing > > > I am not an expert on this but here is my > 2.314567489364532 cents. > > Well there are different levels/approaches to what > people call DR. Some > believe it is an image restore of the queue manager > at the point of failure. > Others see it as COB (Continuance of Business). The > two views are radically > different from a hardware (and yes! Software) stand > point. > > If you are viewing it as recoverable from a point in > time then all > participants must agree on a point in time. Usually > this is a climb up the > staircase of the tower of babble. If you can get all > the platform people to > sit in a room and agree on a coordination to achieve > this. I will take you > to Vegas for the weekend!!! As you see, I'm not > worried about loosing my > shirt at the tables because to achieve this is > almost impossible. > > The lowest point of view is the COB one. Do you want > to put in place a > framework where you can, with a bit of effort, bring > the business back up to > operating speed to move forward and your business > and IT people will hammer > out what was going on at the time of the disaster > and fix it with their > magic wands, special programs to sync up the servers > and catch isolated > transaction from replaying and lots of bubblegum. > This is usually the > easiest and looks to be what you are talking about. > Remember in a situation > like this the servers aren't in sync so there may be > duplicates. If it's a > financial system you are going to need LOTS of > bubblegum!! > > A third possibility that I failed to mention is a > hardware solution. This > being where you have your OS390 fiber connected to a > redundant site, your > UNIX servers Veritas clustered AND replicated in > SYNC and the NT boxes with > a similar redundancy. I'm not talking about just > mirrored disk arrays here. > IF you run a DR then the boxes failover to their > mirrored replicated > partners and the day moves ahead with minimal > outage. This is EXPENSIVE from > a hardware and support staff point of view. (Someone > has to know how to set > up and monitor this stuff). > > AS I said prior, Hardware, Software or platform > coordination. Pick your > poison, it is all downhill from here with a big > bottle of aspirins waiting > for you at the bottom. > > bobbee > > > > > > > >From: "Anderson, Lizette T. (RyTull)" > <[EMAIL PROTECTED]> > >Reply-To: MQSeries List <[EMAIL PROTECTED]> > >To: [EMAIL PROTECTED] > >Subject: Disaster testing > >Date: Tue, 1 Apr 2003 13:15:03 -0600 > > > >We are in the process of creating a disaster plan > for MQSeries v 5.2 > >running > >on OS/390, Windows 2000 and AS/400. I wanted to > get some ideas. The plan > >is to run a backup of MQSeries on the mainframe > while MQSeries is up. This > >backup would be used to restore MQSeries once we > get to the DR site. The > >server side is handled by another group which means > their backup may have > >been taken at a different time. Maybe even a > different day. This is also > >true of the AS/400. > > > >How are you handling DR testing at your shop? > > > > > >--- Legal Disclaimer: The information contained in > this communication may > >be > >confidential, is intended
Re: How to Monitor queue size full
The product Candle can monitor MQ objects and issue alerts on many platforms.You may visit www.candle.com Thanks. Diwakar. From: "mqm mqm" <[EMAIL PROTECTED]>@AKH-Wien.AC.AT> on 04/03/2003 06:48 AM PST Please respond to "MQSeries List" <[EMAIL PROTECTED]> Sent by: "MQSeries List" <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] cc: Subject: Re: How to Monitor queue size full There's a number of tools that monitor system resource usage or alert when thresholds are reached e.g. BMC Patrol, Tivoli. Any particular platform ? mqm --- "Mittal, Gaurav" <[EMAIL PROTECTED]> wrote: > We have queues defined to have a maximum depth of > 6,40,000 messages and a > max size of 2GB.I know we can monitor the queue > depths using 'High Depth > Event' .Is there a way by which we can monitor the > queue size (in terms of > file size not number of messages) also. > Any pointers would be useful.. > > Thanks n Regds > Gaurav > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Max Channels
All: Background: MQ (5.3.0.1) Server is running on a Win2k Server machine. There is one MQ Q manager for this issue. There are 3 local Qs in the Q manager. The Max channel and max active channels has been set to 1000. A local Win2k Service was used to put message into the qs. Client is Java also on Win2k Server. There are two physical machines with exactly the same Java code (Websphere Application Server). When the app server started, 6 threads are created. 1. Each thread will obtained a Q Manager 2. Loop For Ever (Get Msgs) 2.1: Access a Q from the Q manager 2.2: get msg with the Q 2.3 close the Q. 3. When the Application Server shuts down, the Q manager will be disconnected. Note: If there is a problem during the loop, a reconnect logic will be done to the Q manager. The reconnection is a loop by itself with a interval of 3 sec until I get a good connection. Problem: Due to whatever reason, a reconnection logic is triggered, then we will run out of connection (even I have configured the number to be 1000). Trail Fix: Check the TCP tab's KeepAlive Check Box in the WebSphere MQ Service Snap-in. Change the KeepLiveTime in Registry. (HKEY_ LOCAL_ MACHINE\ SYSTEM\ CurrentControlSet\Services\ Tcpip\ Parameters; KeepAliveTime set to 6). We tested this fix, it seems not working as expected. Any other fix ideas? Thanks. Jerry [EMAIL PROTECTED] Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: XML over HTTP and MQ
> If you do SOAP over MQ does that mean you're not doing > Web Services because you are using a proprietary > protocol (MQ) rather than an open internet standard > protocol (http). Views anyone ? I would not call MQ a protocol. In my view, it's a product. Jan. > > -/ /&Jan van Kemenade \&www.cressida.info -\ Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
Look into the Event Monitor book in the MQ Info Center, you can rely on soultino from there Thanks. Email: [EMAIL PROTECTED] mqm mqm <[EMAIL PROTECTED]To: [EMAIL PROTECTED] O.UK>cc: Sent by: MQSeriesSubject: Re: How to Monitor queue size full List <[EMAIL PROTECTED] N.AC.AT> 04/03/2003 09:48 AM Please respond to MQSeries List There's a number of tools that monitor system resource usage or alert when thresholds are reached e.g. BMC Patrol, Tivoli. Any particular platform ? mqm --- "Mittal, Gaurav" <[EMAIL PROTECTED]> wrote: > We have queues defined to have a maximum depth of > 6,40,000 messages and a > max size of 2GB.I know we can monitor the queue > depths using 'High Depth > Event' .Is there a way by which we can monitor the > queue size (in terms of > file size not number of messages) also. > Any pointers would be useful.. > > Thanks n Regds > Gaurav > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: MQSI promote property question
Title: RE: MQSI promote property question Hi Paul, great insight! How could I fail to come to this simple solution? but it happens too often when one get stuck. many many thanks for your great input. best regards, Ben -Original Message- From: Paul Dix [mailto:[EMAIL PROTECTED]] Sent: Thursday, April 03, 2003 9:49 AM To: [EMAIL PROTECTED] Subject: Re: MQSI promote property question I'd consider sending the exported message flow file through a process that modifies the XML to change the value of FLOWDATA to what you'd like. I don't believe there is anyway to modify what that value is during runtime. If there was, then you would have other options/ways to picking the datasource you want from a pool of them. Anybody know of one? The next release with eclipse will fix this as it provides for a config type file for the deploy where you can make these types of changes. Paul Dix Information Systems - Consultant Northwestern Mutual -Original Message- From: BZhou [SMTP:[EMAIL PROTECTED] Sent: Monday, March 31, 2003 9:11 AM To: MQSERIES Cc: BZhou Subject: MQSI promote property question Hi, we have over 70 msgflows, 50 of them are main flows accessing a common datasource, which is given the name FLOWDATA. This is hardcoding datasource name, which we would like to eliminate, esp because we want to setup more than one system on a server. I know for an hierachical msgflow pattern, this can be easily accomplished by promote property. But given our fragmented msgflow pattern, promote property won't bring any benefit. Considering the majority of MQSI applications don't use hierarchical msgflow pattern, some of you must have already encountered this delimma. I would appreciate if anyone can share his or her insight into this problem. thanks, Benjamin Zhou Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: XML over HTTP and MQ
If you do SOAP over MQ does that mean you're not doing Web Services because you are using a proprietary protocol (MQ) rather than an open internet standard protocol (http). Views anyone ? mqm --- "Heggie, Peter" <[EMAIL PROTECTED]> wrote: > It lacks a bunch of things.. > > XML over HTTP - when you don't care enough to send > the very best.. > > -Original Message- > From: Jason Cornell > [mailto:[EMAIL PROTECTED] > Sent: Wednesday, April 02, 2003 1:13 PM > To: [EMAIL PROTECTED] > Subject: Re: XML over HTTP and MQ > > > HTTP lacks WMQ's "assured" delivery. > > >>> [EMAIL PROTECTED] 04/02/03 08:29AM >>> > Where does or how does MQ fit into the XML over HTTP > space ? Seems like this > method potentially does not need MQ as the > middleware. > > > > Wesley Shaw > OJRP 10th Floor > Work: 804 771 3589 (736-3589) > Pager: 888 436 2805 Mobile 804 512 5260 > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > Instructions for managing your mailing list > subscription are provided in the > Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > This e-mail and any files transmitted with it, are > confidential to National Grid and are intended > solely for the use of the individual or entity to > whom they are addressed. If you have received this > e-mail in error, please contact the National Grid > USA Enterprise Support Center on 508-389-3375 or > 315-428-6360. > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Disaster testing
This is the first real discussion I've seen on this list about DR. My favourite topic ...ha! ha! Everytime you mention DR and MQ to IBM you get pointed to the manual/redbook about restart/recovery. But as Bobbee points out DR is potentially much more than this. Which of Bobbee's three options are people adopting ? Or are you doing something else ? Like just hoping you never have a disaster ? mqm --- "Anderson, Lizette T. (RyTull)" <[EMAIL PROTECTED]> wrote: > Thanks for your response. There goes Vegas. I was > thinking it may be > easier to just start from scratch with new pagesets > and logs. I am not even > sure we can bring MQ down to get a decent backup. > > -Original Message- > From: Robert Broderick > [mailto:[EMAIL PROTECTED] > Sent: Tuesday, April 01, 2003 2:09 PM > To: [EMAIL PROTECTED] > Subject: Re: Disaster testing > > > I am not an expert on this but here is my > 2.314567489364532 cents. > > Well there are different levels/approaches to what > people call DR. Some > believe it is an image restore of the queue manager > at the point of failure. > Others see it as COB (Continuance of Business). The > two views are radically > different from a hardware (and yes! Software) stand > point. > > If you are viewing it as recoverable from a point in > time then all > participants must agree on a point in time. Usually > this is a climb up the > staircase of the tower of babble. If you can get all > the platform people to > sit in a room and agree on a coordination to achieve > this. I will take you > to Vegas for the weekend!!! As you see, I'm not > worried about loosing my > shirt at the tables because to achieve this is > almost impossible. > > The lowest point of view is the COB one. Do you want > to put in place a > framework where you can, with a bit of effort, bring > the business back up to > operating speed to move forward and your business > and IT people will hammer > out what was going on at the time of the disaster > and fix it with their > magic wands, special programs to sync up the servers > and catch isolated > transaction from replaying and lots of bubblegum. > This is usually the > easiest and looks to be what you are talking about. > Remember in a situation > like this the servers aren't in sync so there may be > duplicates. If it's a > financial system you are going to need LOTS of > bubblegum!! > > A third possibility that I failed to mention is a > hardware solution. This > being where you have your OS390 fiber connected to a > redundant site, your > UNIX servers Veritas clustered AND replicated in > SYNC and the NT boxes with > a similar redundancy. I'm not talking about just > mirrored disk arrays here. > IF you run a DR then the boxes failover to their > mirrored replicated > partners and the day moves ahead with minimal > outage. This is EXPENSIVE from > a hardware and support staff point of view. (Someone > has to know how to set > up and monitor this stuff). > > AS I said prior, Hardware, Software or platform > coordination. Pick your > poison, it is all downhill from here with a big > bottle of aspirins waiting > for you at the bottom. > > bobbee > > > > > > > >From: "Anderson, Lizette T. (RyTull)" > <[EMAIL PROTECTED]> > >Reply-To: MQSeries List <[EMAIL PROTECTED]> > >To: [EMAIL PROTECTED] > >Subject: Disaster testing > >Date: Tue, 1 Apr 2003 13:15:03 -0600 > > > >We are in the process of creating a disaster plan > for MQSeries v 5.2 > >running > >on OS/390, Windows 2000 and AS/400. I wanted to > get some ideas. The plan > >is to run a backup of MQSeries on the mainframe > while MQSeries is up. This > >backup would be used to restore MQSeries once we > get to the DR site. The > >server side is handled by another group which means > their backup may have > >been taken at a different time. Maybe even a > different day. This is also > >true of the AS/400. > > > >How are you handling DR testing at your shop? > > > > > >--- Legal Disclaimer: The information contained in > this communication may > >be > >confidential, is intended only for the use of the > recipient named above, > >and > >may be legally privileged. If the reader of this > message is not the > >intended recipient, you are hereby notified that > any dissemination, > >distribution, or copying of this communication, or > any of its contents, is > >strictly prohibited. If you have received this > communication in error, > >please re-send this communication to the sender and > delete the original > >message and any copy of it from your computer > system. Thank you. --- > > > >Instructions for managing your mailing list > subscription are provided in > >the Listserv General Users Guide available at > http://www.lsoft.com > >Archive: http://vm.akh-wien.ac.at/MQSeries.archive > > > _ > The new MSN 8: advanced junk mail protection and 2 > months FREE* > http://join.msn.com/?page=features/junkmail > > Ins
Log Files.
Hi Listers, We have a situation where some messages were "lost". The applications says they did not receive the messages and the vendor said they did send them. I dumped the log files but it ended up in a 2.6G (gigabytes) file. My question: is there a way I can get the messages out of the log files back onto a Q to read them back programmatically? Or a tool that can handle a 2.6G log file. This is NT MQ 5.2. Any help will be appreciated. Thanks Ian
Re: MQSI promote property question
I'd consider sending the exported message flow file through a process that modifies the XML to change the value of FLOWDATA to what you'd like. I don't believe there is anyway to modify what that value is during runtime. If there was, then you would have other options/ways to picking the datasource you want from a pool of them. Anybody know of one? The next release with eclipse will fix this as it provides for a config type file for the deploy where you can make these types of changes. Paul Dix Information Systems - Consultant Northwestern Mutual -Original Message- From: BZhou [SMTP:[EMAIL PROTECTED] Sent: Monday, March 31, 2003 9:11 AM To: MQSERIES Cc: BZhou Subject:MQSI promote property question Hi, we have over 70 msgflows, 50 of them are main flows accessing a common datasource, which is given the name FLOWDATA. This is hardcoding datasource name, which we would like to eliminate, esp because we want to setup more than one system on a server. I know for an hierachical msgflow pattern, this can be easily accomplished by promote property. But given our fragmented msgflow pattern, promote property won't bring any benefit. Considering the majority of MQSI applications don't use hierarchical msgflow pattern, some of you must have already encountered this delimma. I would appreciate if anyone can share his or her insight into this problem. thanks, Benjamin Zhou Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: How to Monitor queue size full
There's a number of tools that monitor system resource usage or alert when thresholds are reached e.g. BMC Patrol, Tivoli. Any particular platform ? mqm --- "Mittal, Gaurav" <[EMAIL PROTECTED]> wrote: > We have queues defined to have a maximum depth of > 6,40,000 messages and a > max size of 2GB.I know we can monitor the queue > depths using 'High Depth > Event' .Is there a way by which we can monitor the > queue size (in terms of > file size not number of messages) also. > Any pointers would be useful.. > > Thanks n Regds > Gaurav > > Instructions for managing your mailing list > subscription are provided in > the Listserv General Users Guide available at > http://www.lsoft.com > Archive: http://vm.akh-wien.ac.at/MQSeries.archive __ Do you Yahoo!? Yahoo! Tax Center - File online, calculators, forms, and more http://tax.yahoo.com Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Flood of AMQ7467/AMQ6378 messages
Bridgette. Thanks for your reply. I am aware of the meaning of these messages, my problem is with the frequency of their appearance in an otherwise idling queue manager. I am trying to figure out what can be causing that many messages to be produced. Heinz Beardsley, Bridgette wrote: Heinz, As I am sure you are aware, AMQ7467 is telling you to clean up your logs. Your client should have procedures in place to maintain the log files, which would prevent this message from being repeated. -Original Message- From: Heinz Klein [mailto:[EMAIL PROTECTED] Sent: Wednesday, April 02, 2003 3:10 PM To: [EMAIL PROTECTED] Subject: Flood of AMQ7467/AMQ6378 messages Hello. I appended a similar question some time ago but at that time the customer did not pursue the subject and I never got a complete picture. What is happening: every 3:25 minutes (approximately 200 seconds) the messages above are written to the AMQERRnn logs. In the logs I am serching right now this went on for about 18 hours, without a single change in the log numbers. I saw in an IBM document that for Unix (actually Sun, but I am allowing myself to extrapolate it to AIX, HP/UX and Linux) there is a stanza in the qm.ini file called 'CheckPointWaitLen' which could explain this behavior. However, I am on Windows 2K. Does a similar parameter exist there? And if yes, where can I find it (a quick search through the registry did not show anything similar). Thanks in advance for any help. Heinz Klein OLTP Tecnologia & Solucoes Sao Paulo/SP - Brasil Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
MQWF
Hello, Does someone know if MQWF v3.2.2 is compatible with DB2 UDB v7.2. Thanks in advance for any help. Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Re: Create NEXTSIBLING WMQI 2.1
Use the following statement: CREATE NEXTSIBLING OF NAME VALUE ; I hope this help you Cheers, Manuel Carlos Rodriguez IBM Certified Specialist - WebSphere MQ > -Mensaje original- > De: Declan Harrington [SMTP:[EMAIL PROTECTED] > Enviado el: Thursday, April 03, 2003 11:44 AM > Para: [EMAIL PROTECTED] > Asunto: Create NEXTSIBLING WMQI 2.1 > > If I want to use references for both input and output in a WHILE loop. > Anyone know how do I create the next node of the output within the loop > > ie: > > CREATE FIELD OutputRoot.XML.A.B ; > DECLARE OUT AS REFERENCE TO OutputRoot.XML.A.B ; > WHILE LASTMOVE (IN) DO > SET OUT.NAME = IN.NAME; > -- MOVE OUT NEXTSIBLING or CREATE NEXTSIBLING OUT > MOVE IN NEXTSIBLING; > END WHILE; > > what is the correct syntax for the create? and when I create it do I then > need to move it? > > > Thanks, > > Declan > > Instructions for managing your mailing list subscription are provided in the Listserv General Users Guide available at http://www.lsoft.com Archive: http://vm.akh-wien.ac.at/MQSeries.archive
Create NEXTSIBLING WMQI 2.1
If I want to use references for both input and output in a WHILE loop. Anyone know how do I create the next node of the output within the loop ie: CREATE FIELD OutputRoot.XML.A.B ; DECLARE OUT AS REFERENCE TO OutputRoot.XML.A.B ; WHILE LASTMOVE (IN) DO SET OUT.NAME = IN.NAME; -- MOVE OUT NEXTSIBLING or CREATE NEXTSIBLING OUT MOVE IN NEXTSIBLING; END WHILE; what is the correct syntax for the create? and when I create it do I then need to move it? Thanks, Declan