Re: HYPERPAVs
Take a look at the RMF IOQ reports to see useful Q length and wait times for HyperPAVs. Regards, John >The devices show the candidates: >HYPERPAV ALIASES CONFIGURED = 64 > >And we're even seeing some of them in use. > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Typical FC frame size for FICON
Hello Radoslaw, I'm not sure I can help but here is some information that may be useful. I typically see different average transmit and receive frame sizes during the day (on-line) compared to the evening (batch). I've seen down to 1000 bytes during on-line and around 1700 bytes during batch for DASD. I haven't actually looked only at tape but would expect the average to be higher and consistent. The ISL average sizes seem to be even more disperse. John. On Wed, 23 Mar 2011 09:05:51 +0100, R.S. wrote: >What is typical FC frame size in FICON environment? >I guess there are two different answers - one for disk, the second for >tape. > > >Background: I would like to estimate number of BB credits needed for ISL >between two FICON switches. > >-- >Radoslaw Skorupka >Lodz, Poland -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Typical FC frame size for FICON
Radoslaw, I work for IntelliMagic and use our Vision tool to produce the charts. The information is available in RMF (FCD record 74.7) but you will probably have to turn on the data creation on your switches first. Actually, the most important data that you should be looking at here is the frame pacing count (you want that to be 0) and the Error count. I don't know the configuration of the data that I am looking at but I would guess that only a portion of the data is going via the ISL. If I look at the total receive/transmit rates for all the directors, the ISLs are definitely doing less. John >John >First, I'd like to thank you for the information, I appreciate it. > >I also have some further questions: >1. How do you measure frame sizez? What tool/monitor you use? > >2. Why ISL traffic is more disperse - that means - different? >Assuming he ISL is not used for any other traffic but FICON, it should >contain exactly the same frames + very small amount of control >information. Switch does not rebuild frames. That's theory ;-) Probably >you said about practice, measured values. > >Regards >-- >Radoslaw Skorupka >Lodz, Poland > > >John Ticic pisze: >> Hello Radoslaw, >> >> I'm not sure I can help but here is some information that may be useful. >> >> I typically see different average transmit and receive frame sizes during >> the day (on-line) compared to the evening (batch). I've seen down to >> 1000 bytes during on-line and around 1700 bytes during batch for DASD. >> I haven't actually looked only at tape but would expect the average to >> be higher and consistent. >> >> The ISL average sizes seem to be even more disperse. >> >> John. >> >> On Wed, 23 Mar 2011 09:05:51 +0100, R.S. >> wrote: >> >>> What is typical FC frame size in FICON environment? >>> I guess there are two different answers - one for disk, the second for >>> tape. >>> >>> >>> Background: I would like to estimate number of BB credits needed for >> ISL >>> between two FICON switches. >>> >>> -- >>> Radoslaw Skorupka >>> Lodz, Poland -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: zHPF and mixed channels
Radoslaw, my understanding is that zHPF is negotiated between z/OS and the DSS CU if all the channels in the path group are FEx2 or FEx4. So, if this is the case, then I/O between that z/OS LPAR and that CU can use zHPF. So, my take on Q1 is that the host that has FEx2/FEx4 will be able to use zHPF and the other host will not be able to. In Q2, the Host can negotiate a zHPF connection to one DASD and not to the other. These are only my expectations and I would be interested in your findings - especially if they are different. SMF 73 (RMF Channel stats) will give you some nice information on how many I/Os are using transport mode. SMF 74.5 (Rank statistics) will give you detailed zHPF information per device (starting with z/OS 1.12) (probably Vendor DSS specific). John >Quotation from some IBM TechDoc: > >"zHPF is supported on FICON Express2 (FEx2) and FICON Express4 (FEx4) >channels. It is not supported on FICON Express channels. >"I/Os executed on path groups with a mix of FICON Express and FICON >Express2 or FICON Express4 channels in the same path group will not be >eligible to convert to zHPF. For zHPF, all the channels in the same >channel path group have to be FEx2 or FEx4. " > >The above is clear in scenario with one host and one CU. >What about many-to-many scenarios? > >Q1: The following scenario: two host connected to single CU, via switch. >Same channel adapter in CU. One host is zHPF capable, the other is not >(FICON Express "1"). Will zHPF work between CU and first host? > >Q2: Another scenario: zHPF-capable host connected via switch to two >different DASD boxes. First DASD is zHPF-capable, the second is not. >Both DASDs are connected to the same CHPIDs. >Will zHPF be used for communication between host and first CU? > >Any clue? > >Regards >-- >Radoslaw Skorupka >Lodz, Poland -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Consolidate Storage Groups
>One question I haven't found the answer to: Is there anything special I have >to worry about with the DB2 storage groups and/or classes. I know DB2 has a >thing where it defines one or the other (not sure which) but is there any >connection between the DB2 definition/setting and the SMS configuration? >Will the DBA's need to make changes to reflect the changes in the SMS >configuration? > As long as your SMS routines still assign DB2 data sets to storage group (s) you should need any other changes. >My plan for merging storage groups is to: Looks good. > >For storage classes: >1. Modify the ACS STORCLAS routine so that everything assigned to a storage >class being merged is directed to the target storage class. >2. Will not actually delete the merging storage classes from the SMS >configuration or the STORGRP routine. >3. Eventually move datasets that don't get redefined with the new storage >class. >4. Once there are no longer datasets associated with the old storage classes, >delete them from both the configuration and the ACS STORGRP routine. > >Did I miss something important? > Storage classes are a little different. If someone specifies a particular storage class in JCL to stripe a data set or select a specific volume (guarateed space), you'll have to ensure that you allow for that in your new configuration. You may decide to consolidate management classes or data classes in the future. Same issue as with storage classes; they may be specified in JCL. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SMF74PSM - PAV Samples
Chris, useful HyperPAV activity can be observed at the LCU level in the 78.3 records. But it sounds like you're trying to break down the HyperPAV activity to the device level. Are you using fields SMF74NUX and SMF74PSM to quantify alias assignement? Field SMF74QUE (IOSQ) is really your indicator of adequate aliases. John >Thanks for the info Luis. I guess I was looking for a way to quantify the number of HyperPAV aliases assigned to a given base. For instance, if RMF showed an average of 1 during a one minute interval and the PAV samples were every second then that base had 60 aliases assigned to it. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Calculate Tape Bytes to Tracks
>> I'm curious how you might be expecting to factor in IDRC compression with >> the data stored on the tape? I believe that the BLKCNT represents what is >> being stored, not what got sent down the channel. >> >On tape drives with IDRC or other compression, MVS is only aware of >logical blocks sent across the channel, not the physical representation >on the tape (except very indirectly by indicators of % of media used). >My understanding is that with compression, the compressed logical blocks >are assembled by the tape controller into "super blocks" that are >written on the physical tape, but that structure is not communicated >back to MVS because those are issues that are completely handled at the >tape controller level. > SMF 21 now contains compression information (see APAR OA20077). Regards John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: How to analyze a volume's access by dataset
>Cast your mind back to GTFPARS. This IBM FDP would build seek histograms >using IEHLIST VTOC Listings as input to map the extents of the datasets on >the volume. > Ah Ron, the good old days. One of my tasks at NAB was using GTFPARs to map out the SYSRES access pattern and build allocation JCL to optimize data set placement around the VTOC. Regards John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Calculating the pipe size for DASD mirroring
>If you want a complete independent source with top notch tools and >expertise go to the one many of the storage companies do Intellimagic. > >http://www.intellimagic.net/en/product.phtml?p=Copy%20Services > >We license some of their tooling directly now and have had good >experience with the tools and the available expertise. > >Getting this wrong is can be very painful and upset your cost model for >a proposal so IMHO it is worthwhile to engage with en expert either >allied with your storage vendor of choice or independent. > > >>> >Out of curiosity, how do they stack up against Dr H Pat and his pals over >at _www.perfassoc.com_ (http://www.perfassoc.com) ? > Dr. Pat Artis (www.perfassoc.com) and Dr. Gilbert Houtekamer (www.intellimagic.net) wrote the book titled 'MVS I/O Subsystems: Configuration Management and Performance Analysis' together. Alan, you've already had some good answers and there is a wealth of information available in RMF that can help you calculate throughput. Two of the easiest ways (if you have the right hardware) is to look at the LINK statistics (SMF 74.8) or the FICON switch port values (74.7). Both of these records will give you the write MB/s per link or port. If you intend mirroring all write activity, you now simply have to find the largest interval (peak) and add a safety margin. There are still various link replication details (compression, etc.) but you already have a good starting point. Note: The highest I/O rate typically does not correspond to the interval with the highest throughput. For some technical reading, look at our whitepapers (http://www.intellimagic.net/en/doc.phtml?p=Whitepapers) where we go into detail on obtaining information for sizing synchronous and asynchronous replication. John (alias john.ti...@intellimagic.net) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IBM FLASHCOPY interface ( is there a no cost solution)
If you feel like doing some coding, you can use the ANTRQST API. Take a look at z/OS DFSMSdfp Advanced Copy Services (SC25-0428) for more details. Starting with z/OS 1.12, the REXX interface to ANTRQST will be documented. If you want a preview of how to code the REXX calls, take a look at SYS1.DGTCLIB(ANTFREXX) for a sample. John >DFSMSdss provides all the interface for all the FCV2 functions you may want >to use. Is there a specific function that you cannot get find in the >DFSMSdss support? > >Did you know can use the COPY command in the ISMF dataset list to build and >submit a DFSMSdss batch job. It just doesn't have a foreground option. >Perhaps the FlashCopy processing you require is in those panels. > >> All, >> >> We have migrated our DASD and our new vendor utilized FLASHCOPY. >> >> >> We have CA-DISK & DFDSS only. >> >> >> I've been using TSO =6 and even Batch TSO. >> >> >> Is there any ISPF panels or REXX freebies available ? >> -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: GRS outbound requests
Take a look at the GRS monitoring tool (MVS Planning: Global Resource Serialization - SA22-7600). It will give you an insight into what GRS is doing. (GRS may not be your problem!) John >We have a 6 way sysplex. I'm looking at some performance issues for XCF >traffic. When I look at the outbound requests, I notice that every hour (15 >min interval) group SYSGRS (all members), SYSENF (only one member, CNS >system) and IXCLO012 (all members) has a very high rate of outbound >requests. It's between 2-6 times more outbound req. than the other three 15- >min intervals on an hourly basis. Every hour looks the same. I have looked at >this forum and some other manuals/forums, but don't find a clear answer. GRS >doing something, application related? Any idea or related >documentation I can look at. >/Tomas -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: SMS overflow storage group question
An overflow storage group is used when there is not enough available space for the initial allocation. I believe it doesn't come into the picture when extending data sets. Volumes residing in overflow storage groups are preferred over quiesced volumes and storage groups. Look at DFSMS Storage Administrator Reference (SC26-7402) for a description of allocation (Conventional Volume Selection chapter). Also look at the SHARE website for a presentation on volume selection (3043 - How SMS Volume Selection Works) John >In our general SMS storage group we have several volumes in disabled new >status that the operators can change to quiesced new status if >necessary. What I want to do is set up a pool of volumes that will only >be used if absolutely no other volumes have the necessary space. > >Would an overflow storage group with quiesced volumes be the answer? > >What would be the difference between that just a quiesced storage group? > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: "A foolish consistancy" or "3390 cyl/track architecture"
Eric, the SUN/STK Virtual Tape system (VSM) uses this kind of DASD as their internal buffer (SVA). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
IBM Journals availability
Hello, it looks like one of my favorite sources of information is no longer available for free. Access to the IBM Journal of Research and Development is now only available for a subscription fee (the Systems Journal has been integrated into the Research Journal). >The IBM journals have changed and are currently only available online for a fee. Institutions can subscribe for just $995.00/yr. Not a good move :-( John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Validating existance of a Volume from REXX
Instead of using D U, try using DS. This will not get you the VOLSER for an offline device, but if you know the device address, you can check if it exists in the DSS and actually get some device characteristics from it (i.e. # of cyl). E.g. DS QD,0A80,1,RDC will work for any device number and it will let you know if you have a UCB defined, if it is online (with VOLSER) or offline and how many cylinders. John >If the volume is offline - it could be in any number of "funny" states - maybe halfway thru an ICKDSF init or clip - what would your program want to do with the data in this case? > >Unless you have a real pressing need to determine the volser (if any) of an offline volume (and lets be honest there are not very many reasons for this), then I would suggest forgetting about the offline volumes. > >> >>How to deal with a volume that is valid but offline? The D U fail because the volume is offline. This is one issue I have been struggling with. >> >>How can I see the volumes when they are offline in REXX or through a Batch interface? I know that ISMF can see volumes offline, but I have not found a way to emulate that outside of ISMF yet. >> >>Lizette >> -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CFW with Syncsort
> Yes, SyncSort does use CFW. I wrote the code myself in a previous life, > 17 or so years ago. However, IIRC, there may be circumstances where it is > turned off due to other performance considerations. Check with SyncSort's > Customer Service if you need to know when, if ever these days. > >Hi all, hopefully an easy one here. > > > >Can anyone confirm whether or not Syncsort uses CFW? I know DF/Sort does > >but I'm not sure about Sync. Hyperswap would be one of those circumstances were you would also want the I/O to be written to the disk. In that case you must ensure that CFW is not being used. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CFW with Syncsort
> John, > > I have been told that CFW must be turned off for Hyperswap, but I never > understood why. You say that CFW must be written to disk for Hyperswap, > but > why can all the other writes still be deferred in cache/NVS? > > Ron > > > > > Hyperswap would be one of those circumstances were you would also want > > the > > I/O to be written to the disk. In that case you must ensure that CFW > is > > not > > being used. > > > > John > > G'day Ron, sorry, I hadn't seen your update. 'written to the disk' wasn't exactly what I meant to say! 'All writes have to be in the secondary disk subsystem before the I/O is completed' is more accurate. This is not the case for CFW. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CFW with Syncsort
> > John, > > > > I have been told that CFW must be turned off for Hyperswap, but I never > > understood why. You say that CFW must be written to disk for Hyperswap, > > but > > why can all the other writes still be deferred in cache/NVS? > > > > Ron > > > > > > > > Hyperswap would be one of those circumstances were you would also want > > > the > > > I/O to be written to the disk. In that case you must ensure that CFW > > is > > > not > > > being used. > > > > > > John > > > > > G'day Ron, > > sorry, I hadn't seen your update. > > 'written to the disk' wasn't exactly what I meant to say! > > 'All writes have to be in the secondary disk subsystem before the I/O is > completed' is more accurate. This is not the case for CFW. > > John G'day Ron, I should know better than to argue with you! You're right. It looks like the GDPS group also misunderstood CFW. Excerpt from DS8000 introduction and planning. cache fast write A form of the fast-write operation in which the storage server writes the data directly to cache, where it is available for later destaging. Excerpt from GDPS You should eliminate any known exploitation of Cache Fast Write (CFW). Disk write operations using CFW (Cache Fast Write) are written into cache but not to the disk. Having CFW operations in progress at the time a HyperSwap occurs can yield unpredictable results since there is no corresponding mirrored cache content in the secondary disk subsystem. Any known exploiters of CFW (such as DFSORT) should be customized to not use CFW. And just to complete the thread. To turn CFW on or off at the LCU level use IDCAMS SETCACHE. --- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Where to find CPU% Used by an Address Space
-- snip -- Can someone please point me to a way to find out what percentage of the machine an address space is using? It could be percentage of CPU also and then I can find number of CPU's and do the math, or total percentage depending on how the numbers are stored. I'm not sure how to do this. . . . It's for a simple Rexx exec report I'm writing. -- snip -- Lindy, take a look at the OUCB control block. If you want to get performance data. RMF (assuming you have it running), has already done all the hard work for you. Look at the ERBSMFI interface (RMF manual) for easy access to the data from REXX. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Where to find CPU% Used by an Address Space
-- snip --
Does anyone have any Rexx that calls ERBSMFI directly?
-- snip --
For starters (DEVICE section - but basically the same for other sections)
/*
Use RMF and SMF79 record to get volume information
ERBSMFI requires that RMF monitor I is up and running
*/
rmf_data : nop
bufsize= 90/* Size of SMF record buffer */
smfi_buf = copies(' ',bufsize) /* Initialize buffer */
smfi_req = D2C(1,4) /* Binary: request type=1 */
smfi_rec = D2C(79,4) /* Binary: record type=79 */
smfi_bufl = D2C(bufsize,4)/* Binary: buffer length */
smfi_sub = D2C(9,4) /* Subtype 9 */
smfi_cpu = D2C(9,4) /* Init cpu utilization */
smfi_pag = D2C(9,4) /* Init paging rate */
smfi_parm = D2C(9,2)!!':'
Address "LINKPGM" "ERBSMFI" ,
"smfi_req smfi_rec smfi_sub smfi_buf smfi_bufl smfi_cpu smfi_pag "
smfi_rc = Rc /* Save return code */
if rc <> 0 /* most likely no RMF I running */
then do
/*
rc = 32 means that RMF is not running. Start RMF Monitor I
rc = 28 means the smf buffer was too small. Increase bufsize
*/
smfrc = smfi_rc
exit (smfi_rc)
end
/* Device data section*/
data_sect = X2D(C2X(substr(smfi_buf,45,4)))
data_len = X2D(C2X(substr(smfi_buf,49,2)))
data_numb = X2D(C2X(substr(smfi_buf,51,2)))
loop_start = data_sect + 1
loop_end = loop_start + (data_len * data_numb)
--
John
--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Slip Trap turning off GTF on different LPAR
> I wrote: (somewhat facetiously) > > /RO the_system_you_wanted,the_command_you_wanted? > > And Ed said > Are you suggesting enhancing SLIP to issue system commands? Not at all. I was responding (tongue in cheek) to Peter's comment that there was no system-provided means of driving SLIP on another system. -- snip -- Actually, I like the idea of enhancing SLIP to generate system commands. Then a SLIP trap match could not only stop GTF trace records from being generated, but also stop the GTF STC. Does this sound like a reasonable requirement? -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Installing/Updating SVC
-- snip -- I don't know about SVC 109, but if you want to add a type 3 user SVC (200-255), you can use SETPROG command. For example : SETPROG LPA,ADD,MODNAME=,DSN=USER.LPALIB,SVCNUMDEC=nnn where nnn is 200-255. -- snip -- z/OS 1.9 says - Note: Dynamic LPA does not update system control blocks, which includes the SVC table. To add or replace an SVC routine, the user of the SETPROG LPA command (or the owner of the SVC) must first write a program that gets control as the CSVDYLPA exit routine, looks for a given routine name, and issues the appropriate SVCUPDTE service call to update to the SVC table. For more information on the CSVDYLPA exit routine, see "Monitoring Dynamic LPA Processing" in z/OS MVS Programming: Authorized Assembler Services Guide, SA22-7608. Where did you get the SVCNUMDEC=nnn from? -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Acceptable Syslog Policy(was Determine last LLA Refresh)
-- snip -- My point is most companies use HSM *or* competitive product. So any 'HSM' (as a general description of product class) is present. Not because of syslog archivization, but for backup&ML2 purposes. Only small shops doesn't use it - am I rights ? Now, we have HSM, so my question is why to pay bucks for another software. What are the reasons to pay for it. Caution: I'm not trying to compare HSM, ABR and others. I'm trying to understand why to pay for something more already having HSM-or-FDR-or-others. -- snip -- Radoslaw, let me give you an example (not restricted to SYSLOG output). You do a lot of SMP/E work and would like to keep the output (for what ever reason!). Without a spool archival product you have to do it yourself. You either plan early and direct your SMP output to data sets in the JCL, or you use the SDSF/IOF options to copy your spool output to DASD (you could also use spool offload), give it a reasonable SMS management class and let HSM/FDR manage it. Unfortunately, in real life there are 500 developers and 8000 production jobs and many tasks generating a lot of spool output that may need archiving. Doing it manually is no longer reasonable. That is why this kind of software is very valuable. Grouping spool output together, restricting access, managing output (with different retention requirements) can all be taken care of by one piece of software. HSM/FDR can't do that. If you are only talking about SYSLOG output then I agree that you can just pull it off spool and let HSM/FDR manage the data sets. -- snip -- BTW: What are the "others" ? CA-DISK ? -- snip -- There are quite a few of these products. BETA 92 comes to mind. If you don't want to spend money, there is an output management program on the CBT tape. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IBM paper on Large Stand Alone Dump handling
-- snip -- 1. Ontop for EMEA customers: I think ontop is dead in EMEA. IBM consistently refused to use ontop to look at a dump, always demanding the dump to be ftp'd in. That was (and still is) a nuisance. A colleague then wrote a program that allows us to ftp directly to IBM (don't ask me how), but without it we would still send cartridges. We have now deleted the environment. -- snip -- I prefer the ftp method to ontop. Once you set up a job to terse and send the dump to ftp.emea.inm.com it works quite well. With the old ontop method, there were constant problems with the user_id passwords and racf access. -- snip -- 2. I have just done some testing for standalone dump to a large dasd data set (and fallen flat on my face while I was at it). A word of warning: The UCB address(es) of the DASD data set are put into the preallocated DASD data set. Make sure they don't change! (This may be what the paper refers to as hiperswap, and I don't know what it is.) We mirror our DASD and when we swapped, the ucb address for each volume changed to its mirror while the former mirror became the primary. This failed sadump spectacularly! (with a message that didn't tell me anything under 1.8. I am told it changed to something meaningful with 1.9) We have now moved that data set to non-mirrored DASD. -- snip -- Hyperswap is one of the best things that have happened to reduce DASD outages. It works very well. But. Whether it was a hyperswap or a normal mirror swap, you have to be aware that the DASD device addresses have changed. GDPS takes care of this when LPARs are IPL'd (primary and secondary IPL information is primed). For SADMP, I would still aim to mirror all DASD and have separate volumes prepared for SADMP usage on the primary and SADMP usage on the secondary side. Set up separate procedures for starting it depending on whether you are running on the primary or on the secondary DASD (you always need to be aware of where your DASD currently is!). If you hardcode only a primary device number for SADMP, what do you use when your primary DASD is really hosed? -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: REGION=0M and LSQA
-- snip -- > (expecting a flame or two :-) Well, I type too slowly to do much flaming. And of course, I am willing to discuss this in great detail any time you come to Poughkeepsie and take me out for a beer. However, I am at somewhat of a loss to figure out how an operating system would distinguish between a program which intentionally uses a lot of storage, and a "runaway" program which unintentionally uses a lot of storage, without someone telling the operating system how much storage the program is intended to use. -- snip -- My perfect world. No artificial limits set, only the architectural, and z/OS controls each task individually, according to what the current (storage) environment allows for. z/OS knows what the limits are and can work out how well the ASM infrastructure is set up to cope with paging at that moment. WLM has classified tasks, and z/OS could certainly allow higher valued tasks a much faster storage growth than the lower classified tasks. The action that is initiated when a storage shortage has occurred could also be different for the different types of tasks. A program using a lot of storage would not need to be judged as intentionally using the storage or a "runaway getmain" task. If the task storage usage gets so large that the z/OS image is endangered (ASM shortage or other paging problems), then z/OS needs to react quickly enough to isolate the problem. The decision on whether it is a cancel or a quiesce (or something else - like compressing the pages to get rid of repetitive data and paging it out!), could be decided based on WLM categorization. Jim, I'll take you up on that beer offer - in Cologne :) -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Exploiting Parallel Access Volumes
-- snip -- My understanding of channel program execution is that there is one queue per UCB, i.e per volume. In other words there isn't a huge amount to be gained by executing lots of concurent channel programs to the same file (you'd get some advantage, maybe, in that you'd be able to construct your next EXCP invocation and have the EXCP do all the prep work for it before actually checking that you previous one is done). Parallel access volumes however mean that there are potentially multiple UCBs per device and therefore multiple channel program queues. However if I examine the list of UCB associated with the extents of my file they are all the same. So is(are) parallel access volume technology something I can exploit in my application? The best I could presumably do is to allow multiple channel programs to be scheduled by my application against different extents in the file, but since my UCB addresses are all the same, that's not going to work. Is this an MVS configuration issue? i.e is there something that would need to be set up on the machine to allow this? Or is this more a case of different jobs will get different UCB addresses for the same volume, so they can be parallelized, but you can't exploit that within a single app? Searching on the web, Syncsort reckons it exploits PAV to get 30 percent improvement in elapsed time or something. I want a piece of that if I can get it. -- snip -- Well, let me give you my take on what it looks like at a higher level than your EXCP programming. I'm sure someone will jump in and correct me if I'm wrong. The UCB is the z/OS view of the device for the application program. It stays that way even when PAVs are implemented. In the hardware definition, this UCB would correspond to a base address. The PAV addresses are defined in the hardware as aliases. PAVs can currently be implemented in 3 ways. Static, WLM managed and HyperPAV. A static definition (ALIAS to BASE mapping) is defined in the hardware. WLM managed PAVs get moved around according to I/O load (the ALIAS to BASE mapping changes). HyperPAVs get assigned to every new I/O and are managed by the hardware. The DS QD,x,y,UCB command shows me the UCB for a device. This only shows me a UCB for the BASE address (It doesn't work for an ALIAS address). The UCBFL indicates whether the UCB is a BASE, ALIAS or HyperPAV, so maybe the DS command is restricted to only base UCBs. The DS QP,x,y shows me information for the base or the alias address. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: S30A-10 Abend and S878-18 abend during freemain
> Our shop had migrated to z/os 1.7 from z/OS 1.4. > > After migration one of our production job is abending with S30A-10 > abend during freemain and subsequently the job encounters another > abend S878-18 (with z/OS 1.7). > > When we resubmitted the same job in another two different LPAR that > has z/OS 1.4 the same job runs successfully. > > We then commented out the freemain code and resubmitted the test job > again in z/os 1.7. We did this so that the operating system would do > the freemain for us. But even this time we got the same S-30A abend > and then S-878 abend. > > We are struck with this issue for a long time now. Really don't know > how to proceed further on this. Any help on this issue will be great > help for us. Have you taken a dump and looked at the FREEMAIN in question (and maybe the VSMDATA part as well). S30A-10 indicates that you're trying to freemain LSQA. Is this what you want to do? John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: S30A-10 Abend and S878-18 abend during freemain
> > Our shop had migrated to z/os 1.7 from z/OS 1.4. > > > After migration one of our production job is abending with S30A-10 > > abend during freemain and subsequently the job encounters another > > abend S878-18 (with z/OS 1.7). > > > When we resubmitted the same job in another two different LPAR that > > has z/OS 1.4 the same job runs successfully. > > > We then commented out the freemain code and resubmitted the test job > > again in z/os 1.7. We did this so that the operating system would do > > the freemain for us. But even this time we got the same S-30A abend > > and then S-878 abend. > > > We are struck with this issue for a long time now. Really don't know > > how to proceed further on this. Any help on this issue will be great > > help for us. > > Have you taken a dump and looked at the FREEMAIN in question (and maybe the > VSMDATA part as well). > > S30A-10 indicates that you're trying to freemain LSQA. Is this what you want > to do? > > John > Hi John, A dump didn't get produced for this abend as we also got S878 abend. But the intention of freemain is to just to release the work area used by the program. One thing which we are not able to understand is why does it fail on the new z/OS 1.7 alone, while the same job has been running successfully over many many years now. -- -- snip -- I would still take a dump (set a SLIP for S30A if you have to). Is the environment the same for your program (APF status of the library the same in both cases)? --- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: S30A-10 Abend and S878-18 abend during freemain
> Hi All, > Thanks for the responses. Apologies for late response. > > We had set a slip trap for S30A abend. We also recompiled and link- > edited all utility routines and still we got the same S30A abend. > > We had then sent the dump to IBM. IBM got back to us saying one of our > inhouse written routine used to sort contents in a memory location was > somehow overlaying the storage area where DCBs of all datasets were > present. This inhouse routine was overlaying just 4 bytes of memory in > the DCB declarations. We just commented out the calls to the inhouse > sort routine and then used an inhouse sort macro to sort the data in > memory. That fixed the S30A abend for us. > > Meanwhile, we also found that replacing SYSOUT=* with a temporary > dataset or a permanent dataset also did on give any S30A abends with z/ > OS v1.7 (strange), without making any changes to code. > > One thing which we still dont understand is that the same job when > supplied with a temporary dataset/permanent dataset runs perfectly > alright with z/os 1.4 and 1.7. Where as it fails with SYSOUT=* in z/os > 1.7 and runs fine in z/os 1.4. > Is the SWA above/below setting the same for both systems? Is it possible that the private area size on z/OS 1.7 is different and your routine just happened to stomp on the LSQA because the gap from your allocated area to LSQA is now smaller? These are only WAGs. Now you know which routine was doing the overlaying - that gives you something definite to work on. > Also, the inhouse program was overlaying storage area contents present > in subpool 0 but the inhouse routine used was only sorting the > contents present in subpool 1. Is it possible for a program to corrupt/ > overlay contents present in different subpools? Look at MVS: Diagnosis for a description of subpools and storage keys. Also, you should register with the list server and send you mails to IBM-MAIN. You'll reach many more fine individuals that can help you in the future. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Restartable PDSE Address space
> >> Is there anyway to create a latch problem on a test PDSE to test the > >> restartable STC? > > > >Ok, this is a WAG. No warranties express or implied... yada yada. > > > >The GRS latch services do not provide any recovery or resource > >management of their own. It is left up to the application, SO... you > >could try doing things that would get you ownership of a latch and then > >perhaps not get it freed. > > > >For example, in an ISPF session, open a PDSE member for edit. Then > >cancel the address space while you're in edit mode. Maybe the latch > >get freed, maybe not. But there's a decent chance the PDSE developers > >covered that case, so if you have a monitor that allows you to abruptly > >MEMTERM an address space (e.g. the MV/Resolve exit command) then you > >could try the same experiment using that instead of using the cancel > >command. That may bypass task level resource managers and trip the case > >you're interested in testing. > > > >Or not. > > > > > And make sure you try this on your most critical LPAR during peak period. > That will give it the best change to fail since there will be a lot of > other activity on the system. :-) > > (I hope you know I am JUST KIDDING!!) > > BTW, even if you did try this in a sandbox and didn't care about potentially > needing an IPL... it would probably be a good idea to have a backup copy > of the PDSE you are testing with. :-) If you can edit the PDSE outside of your sysplex (definitely not recommended!) you could probably create some problems. I suspect that the PDSE address space would not have a problem but the data set itself though. Does the IPCS VERBX SMSXDATA show you PDSE latches? You could just casually free one of these (Vary SMS,.. FREELATCH) and see if it causes any problems (especially if you try to edit the same member from a different address space). I seriously don't think you need to be concerned about the PDSE address space recovery. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CHPID online requires 10-15 min after upgrade to z9
> Is there anyone try "CF CHP(XX),ONLINE" a FIOCN channel requires 10-15 mins > after migrated to z9. Any shop has met before ?? any solution? > > Any comment will be appreciated I haven't seen that kind of a response on a z9. After you've issued the CF CHP command, try issuing a D GRS,C command (a few times) to see if there is any contention. During the 10-15 minutes, do you see any messages on the HMC? -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Is anyone out there using HSM and the CDS are managed via RLS?
-- snip -- Our shop is Z/OS 1.8. We want to implement HSMplex(4 members).we will also implement DR in the future. We are not sure whethere we need RLS for accessing the control data sets. Could you tell us the advantage and disadvantage of RLS ? Is there any limition for the HSM CDS managed via RLS? -- snip -- Jason, my customer has been using VSAM RLS for the HSM CDSs for quite a while. There were problems initially with SMSVSAM, so it's a good idea to be current on maintenance, Make sure that any products that interogate the HSM CDSs directly are checked. One such product was Cleanup. I'm not sure if that product was only available in Europe, but some of the functions no longer work when you convert to RLS serialization. Also make sure that you know how to return back from using RLS in case there are any problems. Some of the SMSVSAM parameters are unique! John. PS. This message came to the list 3 times!! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CACHEFASTWRITE
-- snip -- can you pls tell me if IDCAMS/SETCACHE VOLUME(xx) CACHEFASTWRITE OFF works on a ssid basis or at "storage cabinet" basis meaning when I run CFW OFF for one volser I get all devices from the storage cabinet with CFW OFF. -- snip -- It gets set per volume. If you want all volumes to have CFW turned off, then you have to execute the command for each of them. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CACHEFASTWRITE
-- snip -- John has answered your question, but I would ask that you consult your storage vendor about the recommendations for this value. HDS for example generally recommend having CFW turned on as it can halve the amount of cache used for CFW writes that have not been destaged to disk. -- snip -- Ron, when HYPERSWAP is implemented, IBM requires that CFW be turned off. It could be the reason that Antonio is asking. Regards John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CACHEFASTWRITE
-- snip -- As far as I can see the CFW commands always apply to the whole subsystem, and have never applied to a single volume. My past experience is that subsystem means the whole Storage Controller. My recent experience is with HDS only and this is how the command operates still. -- snip -- My mistake Ron. The CFW command does indeed apply to the whole subsystem (also true for IBM ESS and DSx000). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ASSIGNED TO ANOTHER SYSTEM (IEE791I)
-- snip -- If the unit is being used in a sysplex then the following command issued on one system should suffice : RO *ALL,D U,,,addr However, as Shane points out if the unit is online to a system outside the sysplex, then I cannot think of a better procedure than : (a) Gulp (b) Issue the "D U,,,addr" command on each system until found. (c) Do something to ensure that this doesn't happen again (HCD, auto-ops, configxx etc etc) -- snip -- I used to work for a client that shared tape drives between IBM and Siemens hardware. Keeping a tab on which system currently has a tape assigned is possible with a little work. Rob's (b) suggestion could include some kind of a procedure that sends the Display Unit commands via TCP/IP to the different sysplexes (or OEM systems) and then returns the results. Not trivial - just an idea. The MVS command DS QTAPE can be used to return device information - but unfortunately, the assigned information is not available. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Virtual tape limits (Was: OEM software electronic download report card)
-- snip -- Just ran into that. Things that large should be on physical tape (if you have it). We ran a report of dsns with more than 20 volsers and are just about all of them were from DB2 and jobs created by the same DBA. Our default forces things to virtual and people do have to let us know if they want it to go to physical tape. -- snip -- VSM (and VTS) offer simple, application independent duplication for disaster recovery purposes. When you change jobs from using virtual tapes to real tapes you have to look at how you're going to duplicate the tapes. It's one reason to try to keep all your tapes virtual. -- snip -- As far as more options... I know you use Sun/STK like we do... I think I heard VTCS 6.2 will have some help there. -- snip -- As has already been indicated elsewhere, the size of the virtual tapes has been increased. There will still be data sets that are too large, but it will make it better. BTW. What are your plans (Shane/Mark) for migrating to VTCS 6.2. Specifically, how long do you plan to wait before installing the new software? John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Channel Detected errors only 1 lpar, 1 job, 1 vsm
-- snip -- Alright this is a head scratcher at least for me... We have recently installed a VSM5 and also have some VSM4's in the mix. We have one job that runs that is continually getting channel detected errors every time it runs and the errors are only occurring on the VSM5. Below is a small example. It shows one tape mount that works fine, and then the other that gets the error... This is using FDR so perhaps Bruce Black has some insight... . . . IOS050I CHANNEL DETECTED ERROR ON 1891,17,01,**02,PCHID=0231 IOS050I CHANNEL DETECTED ERROR ON 1891,13,01,**02,PCHID=0140 -- snip -- If writing this data set to a VSM4 type works, and doesn't work on the VSM5, then you should open a problem with SUN. You may need to take VTCS and GTF trace data, but the SUN technical people will request what data they need. My customer has a couple of VSM5 systems installed and we do not have the problem you are describing (DFDSS user). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Channel Detected errors only 1 lpar, 1 job, 1 vsm
-- snip -- John, Go to the HMC and do problem analysis on the PCHID to see if you are getting sequence errors. Possibly a bad cable or connection. Doug - Original Message - > IOS050I CHANNEL DETECTED ERROR ON 1891,17,01,**02,PCHID=0231 > IOS050I CHANNEL DETECTED ERROR ON 1891,13,01,**02,PCHID=0140 -- snip -- Doug, the errors occur on two different channels (and PCHIDs). If the VSM5 box is connected to a director or a switch of some kind, then I suppose that the cable between switch and VSM5 could be causing the problems. The originator of this thread didn't say whether he'd tried all paths, or how the box is connected to his server (I was going to write CEC !!). John (not the originator) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Virtual tape limits (Was: OEM software electronic download report card)
-- snip -- > VSM (and VTS) offer simple, application independent duplication for > disaster recovery purposes. When you change jobs from using virtual tapes > to real tapes you have to look at how you're going to duplicate the tapes. > It's one reason to try to keep all your tapes virtual. [...] It's one of the reasons to use HSM or FDR to make duplex copies. It costs no more CPU cycles, it fills up the tapes, it utilizes more channels (it's not a problem IMHO). If you want to write to tape directly... well... WHY do you want it? -- snip-- With most installations it's purely historical. The big data sets went directly to tape, the smaller ones went onto DASD and were then processed by HSM (or similar). With DASD prices now cheaper and still dropping, I agree with you that there shouldn't be any need to write directly to tape. Why VSM/VTS and not HSM. Well, one reason is that handling duplicates in HSM requires manual intervention. When a primary tape goes bad, you need to activate the alternate and then ensure that you produce another duplicate of this tape. All quite simple, but still manual. When a virtual tape is bad (due to a bad real physical tape), it's all handled under the covers as far as HSM is concerned. No need to screw around with primary and alternate. You do need to invest more in the size of your VSM/VTS and maybe that is reason enough not to do this. Also, implementing high availability is a lot easier when the duplexing is application independant. -- snip -- BTW: The biggest VSM/VTS advantage I in my opinion is the number of drives. It's important in multi-LPAR installation. -- snip -- That's true, but it is also important to size your VSM/VTS properly to ensure that residency time is long enough for your virtual tapes and that the tapes get migrated to the backend in a timely matter. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Virtual tape limits (Was: OEM software electronic download report card)
-- snip -- > >BTW: Even without clustering it is not necessary to move MVCs, since you can write VTVs on two MVCs in two ACSes (in two locations). > Correct. We are looking at doing one of the two methods hopefully by the end of this year. Unfortunately at the moment we still work with PTAM (Pickup Truck Access Method). -- snip -- One of the considerations that needs to be looked at when writing tape to two locations is how it is done. The cheaper method is asynchronous, namely, the logical tape gets written to the physical backend media at some time 'in the near future'. If you don't need any synchronisation between DASD and Tape at your remote location, that's OK. Depending on how busy the VSM boxes are, the writing of the logical volume to the physical media could take a while (many hours). Ensure you carefully define your response time objectives and see if you can meet them. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IEBCOPY Unloaded dataset to PC and back again...not
-- snip -- > >FSVO "ALL". Because BINARY probably doesn't work, > > WTF? I have *NEVER* seen a failure in a binary FTP transfer, other > than the usual timeouts and dropped connections.. > > -- Unless done correctly the FTP of some datasets, such as those with Variable or Undefined record formats, will "fail" in that the resulting file on the PC is unusable due to lack of end-of-record indicators. This is a logical failure and not a physical failure. -- snip -- Wouldn't specifying TYPE E and MODE B get around the problem. MODE B specifies block mode which would then preserve the record format (even if it is undefined). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CA-Allocate / SMS
-- snip -- With VAM (Since I AM a Silverback), you can use a DDNAME of VDSDIAGS to trace the allocation thru the selection routines. This helps to see why a dataset didn't go where (I thought) it should have. . . Does SMS have a similar DD? I've used the TEST facility in SMS, but find it wanting. -- snip -- As Sam has already pointed out, there is nothing available in native SMS other than the ACS test option. You could possibly code up some SMS exits (IGDACSDC,IGDACSSC,IGDACSMC) to do some tracing for you. You could also just add a lot of WRITE statements in your ACS routines and restrict them to your userid. John. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: CA-Allocate / SMS
-- snip -- > > As Sam has already pointed out, there is nothing available in native SMS > other than the ACS test option. Oh ye of little faith! Since z/OS 1.7, there is a console command SETSMS VOLSELMSG which will provide summary or detailed info about the SMS allocation process. It can be limited to specific jobs or dataset names. I haven't tried it but it sounds like just what he needs. -- snip-- Thanks for the Info. Bruce. It's amazing what you can find in the manuals if you just look!! SETSMS VOLSELMSG(ON,ALL),JOBNAME(myjob) returns some useful information that can certainly be used for analysing allocation problems. You can get detailed information indicating why particular volumes were not selected for allocation. Good stuff! Another additional method is to use the SMS trace table. SETSMS TRACE(ON),TYPE(-- specify type or ALL) I haven't done this myself, so it could be that the trace table information (formatted with IPCS and SMSDATA), is not useful for us. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: FDRPAS Question
-- snip -- I have a four CPU complex with multiple LPARS on each CEC and DASD from three vendors attached to all CECs. Using FDRPAS to swap a volume that is only online to one volume, but getting fdrw68 message. Anyone else run into that situation? What did you do to get rid of the message? Any suggestions are appreciated. -- snip -- It's possible that you have the device defined in HCD AND have chpids that are connected to the storaqe unit, but the device isn't online to the LPAR (most likely because you defined it that way in the OSCONF). If you DEFINITELY know that this volume is not online, you can EXCLUDE the CPUID in the SWAP job. Be very careful about excluding CPUs if you choose to do this. I've had to do this in the past for a very large multi-cpu/sysplex environment, where the volumes were not online to various LPARS (and also not defined in their OSCONFs), but were physically reachable. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: How to monitor CPU peaks from an exec?
-- snip -- I would like to get a fairly simple report in real time of an online server's CPU usage. I'm especially looking for transaction-like peaks/spikes of CPU time, percentage of CPU used on the machine, and how long these spikes last. (What RMF does basically.) Right now I have an exec to get the elapsed time and total CPU time used for the address space. This is pretty good and gives me much of what I need. I'd like to beef it up a bit if possible. It would be nice if I could get this information from a Rexx exec running in a TSO address space. If possible, could anyone give me a hint as to which control blocks to look into? The exec is easier for me because I can monitor things, like at a customer site, without having to bother their RMF guys who are usually quit busy. If not, of course I can always get the RMF data. -- snip -- Lindy, if you want to do it in your own exec, take a look at the RMF API (ERBSMFI). This is documented in the RMF Programmer's Guide (SC33-7994). RMF monitor I needs to be up and running for some of the sub-types. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Migration from z/800-0B1 to z9BC-703
-- snip -- When we upgraded to the z9 from a z800, the biggest shock was HSA. 1.4 Gig for HSA and no warning it would be that bad ??? I read about its impact, but never imagined it would be so much. Wiped out a whole LPAR I had planned. I concentrated on reviewing a paper called "Moving to fewer and faster CPU's" (or something like that) when we originally went to the z800. That has a good discussion of the concerns at a level my simple mind can understand. I just wanted to ensure I didn't short something I cared a lot about. All-in-all, our upgrade went well. I'm still learning to optimize the configuration, but you'll be able to hide most of your transgressions behind the increased performance. Moving from the z890, you won't have major conversion concerns either. Considering the jump you're making, happiness will abound. -- snip -- Dave, one of the things I almost always forget to check on is whether the increase in real storage that the LPAR sees will result in MVS re-aligning its virtual boundries. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Software only VTS?
snip --- And let's not forget TMM that comes free with z/OS and your friendly neighbourhood Storage Administrator... Ron --- snip -- That's almost true Ron. Some product has to move the data to tape. You still have to pay for HSM. Other than that, I agree with you. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IOS163A
-- snip -- The HMC will probably also have a hardware message for these errors. The message will give you an indication of the Unit Address that issued the problem, but unfortunately the message will not indicate the LCU. Your IBM CE should beable to find out this information for the message which will pinpoint your problem device(s). .. .. -- snip -- Recently I saw the following error: V 3000,ONLINE *IOS163A CHPID A3 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE *IOS163A CHPID 10 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE *IOS163A CHPID A3 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE *IOS163A CHPID 10 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE *IOS163A CHPID 10 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE *IOS163A CHPID A3 ALERT, NO ASSOCIATED SUBCHANNEL FOR DEVICE IEE302I 3000 ONLINE . . Any clue ? -- snip -- Radoslaw, as someone has already said, it most likely indicates a configuration conflict between the hardware and HCD. I've seen problems like this when the base and alias addesses in an ESS DASD didn't match those that were defined in HCD. In addition to taking a look at the HMC, maybe you could take a look at EREP and see if there are any useful logrec records available. My guess is that they will most likely not provide you with any additional information - but it's worth a look anyway. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Dyn Alloc gives 2nd instance of same GDG name
-- snip -- I have a task that allocates a few GDG files during its running instance. I allocated one instance of a GDG and got the G0207V00 extension. 4 seconds later, I allocated another instance (different DDNAME) and got the very same G0207V00. It would appear that the first 207 instance was "deferred.". And that before it could be 'rolled in', we had allocated another instance of the GDG. . . . The same program is doing the allocation in all instances. The program issues a step level exclusive ENQ on a resource that covers... 1. allocation of file. 2. open of the file. 3. first record write of file. I then release the ENQ, to allow other instances in the region to do the same. I have a task that allocates a few GDG files during its running instance. . . . Is there something behind the scenes in SMS that might be tripping me up? Would a non-SMS managed GDG eliminate this problem? Perhaps I should add a delay as step 4 in my comments above to wait say about 3 to 6 seconds or so before allowing the next allocation request to start? -- snip -- I hate it when I find code waiting for a fixed time, hoping it will be long enough for some action to finish. It will always bite you some time in the future. Back to your problem. If you look at the description for IGD107I, it says - IGD107I dsname ROLLED IN, DDNAME=ddname Explanation: At the time of the step ending, the SMS managed generation data set associated with the DDNAME became a permanent part of the generation data group (GDG). I would say that it is working as designed. Can you use different GDGs within the same step? John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: KEY8 CSA: Pro/JCL and Info/XE
-- snip -- > :>Ed Jaffe asks... > :>> Out of curiosity, which software products allocate CSA in keys A- > :>> F? > > :>I know of one major vendor's storage management product that right up > :>until the last time I looked (more than a year ago) ran in key 10 > :>(X'A0') and used a bunch of CSA in key 10. The developers thought it was > :>clever. I didn't. > > While I don't condone it, I don't see the exposure (unless the standard zOS > problem state key mask includes key 10). A V=R (ADDRSPC=REAL) job step is assigned a key from the 10-15 range. -- snip -- Could you guys clear up a few things for me. I've seen (E)CSA key 14 data in dumps. The fine manual states :- -- manual extract -- (referring to keys) 8-9 All V=V problem programs 10-15 V=R problem programs (each protected by a unique protection key) and also - The system assigns real frames upon request from a pool of available real frames, thereby associating virtual addresses with real addresses. Frames are repossessed upon termination of use, when freed by a user, when a user is swapped-out, or when needed to replenish the available pool. While a virtual page occupies a real frame, the page is considered pageable unless specified otherwise as a system page that must be resident in central storage. The system also allocates virtual equals central (V=R) regions upon request by those programs that cannot tolerate dynamic relocation. Such a region is allocated contiguously from a predefined area of central storage and is non-pageable. Programs in this region do run in dynamic address translation (DAT) mode, although real and virtual addresses are equivalent. -- manual extract -- Does that mean that a Job needs to have ADDRSPC=REAL on the job or step card in order to allocate storage in the key range 10-15? Can one assume that any authorised code can use the V=R keys with out specifiying ADDRSPC=REAL? The storage is non-pageable - but not fixed! What does this mean (other than the obvious!). Namely, what happens when storage is configured offline. Are these pages allocated in the preferred storage area? Is it really the case that real and virtual addresses are equivalent for these storage areas? Thanks John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IBMLINK vs Resource Link
-- snip -- Out of curiosity: What is the difference between IBMLINK and Resource Link? Btw. - Our hardware guys use resource link https://www-01.ibm.com/servers/resourcelink/hom03010.nsf?OpenDatabase - For PMR's I use : Software Service Request https://www14.software.ibm.com/webapp/set2/ssr/slprob - For APAR's I use : Technical help Database http://www14.software.ibm.com/webapp/set2/srchBroker/views/srchBroker.js p?rs=112 -- snip -- Resource link is hardware slanted (ie. Customer initiated upgrade) Your two other links are not IBMLINK but the WEB offering for service/support. Some of the functions of IBMLINK are duplicated, but IBMLINK does a lot more (eg. APAR tracking). SIS in IBMLINK gives me more hits for the same search string compared to the Technical help Database. That can make a difference. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: HSM and offsite logging -- An HSM question
-- snip -- I was hoping there was an exit in HSM where one could code dataset patterns like IMSVS.SLDSS.*. Each time HSM goes looking for work in its input queue, the exit would be invoked to scan for a pattern match and select our IMSVS.SLDSS.* dataset activity for processing, in preference to other commands If you were to write a rexx exec to issue the migrate commands then also use the WAIT sub-parameter, I believe that it will push the MIGRATE commands to the front of the queue -- snip-- Starting with z/OS 1.8 you can use the ALTERPRI REQUEST(xxx) HIGH HSM command to adjust the priority on the request queue (options are HIGH or LOW). -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Why is HSA so ridiculous on a z9 at z/OS V1R8?
-- snip -- Maybe if IBM renamed it people would mind less. Perhaps CFFHASRADHCSA Customer Focused Functionality High Availability Super Reliability Allows Dynamic Hardware Changes Storage Area? -- snip -- If they hid it under the covers (internal, customer non-visible memory) we would all be much happier. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: z/OS 1.9 Announcement Letter ENUS 207-175
-- snip -- z/OS 1.9 manuals are online at http://www-03.ibm.com/servers/eserver/zseries/zos/bkserv/ -- snip -- Also take a look at the new Hot Topics issue with Ms. zIIP and Mr. zAAP on the front cover. Lots of good information! -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IPCS and disassembly/LIST addr INSTRUCTION
-- snip -- Seems like it got confused with an imbedded data area and started disassembling "instructions" at odd addresses. -- snip -- The plague of every disassembler. There are some disassemblers that allow you to specify which area is data (non-code), and specify base registers ... But, for the purpose of quickly looking at a piece of a dump, I find it quite usable. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IPCS and disassembly/LIST addr INSTRUCTION
-- snip -- :>-- snip -- :>Seems like it got confused with an imbedded data area and started :>disassembling "instructions" at odd addresses. :>-- snip -- :>The plague of every disassembler. Not on the S360+ platform. Instructions CANNOT be at odd addresses. Sheesh. -- snip -- Sorry Binyamin, I interpreted your 'odd addresses' differently to the way you meant. Yes, of course you're right. You saw it first; I'll let you open the PMR. :-) -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: JCL passing parms to ASM module?
-- snip -- The OPEN broke it! . . I don't believe I have any entry housekeeping that could break R1: BALR 3,0 USING *,3 OPEN (SYSPRINT,OUTPUT) L 2,0(1) *copy parm addr into R2 etc... -- snip -- Taking a look at the assembler output (using PRINT GEN), will show you the macro expansion. It then becomes apparent what happen to R1 before you tried to copy the parm. -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Protecting ISMF Functions
-- snip -- With this in mind, is there something else protecting us from someone bypassing RACF program control with an unprotected copy of ISPF? Do the STGADMIN.** profiles in the FACILITY class protect the underlying functions? If so, is the RACF program control suggestion offer only a redundant protection? Or am I missing something else? -- snip -- RACF program protection can not only restrict the program call but also restrict the call to a specific library. I'm not sure I would call it redundant. For many users it may be enough protection. The STGADMIN. profiles offer a much finer level of protection - independent of where the ISMF load modules come from. It is what I always set up (especially to stop users making themselves into storage administrators). -- John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: IODF and logical paths limitation
-- snip -- Background: when you define i.e. DASD unit in HCD you can have multiple logical control units (CUADD) as well as many LPARs can use the chpids attached to the CU. However logical paths limit (# of logical paths = #CUADDs x #LPARs) vary depending on DASD model. For example HDS 7700E aka Tetragon 2100, escon-attached, have 16 logical paths per chpid. ESS Shark has 64 paths. -- snip -- 64 Paths are for ESCON attached channels. -- snip -- Questions: 1. Where can I find such numbers ? Usually technical specifications does not cover this parameter or it is "well hidden". -- snip -- You need to know the 'special handshake', then you may get some information from your CE. :-) For the IBM CU, take a look at the redbook IBM System storage DS8000 Series : Architecture and implementation SG24-6786- Comparisson values for ESS are also included. -- snip -- 2. What is the limit for ESS Shark *ficon-attached* ? --> 256 3. What is the limit for HDS 9960 (escon and ficon) ? --> ? 4. What is the limit for EMC DMX-3 ? Is it 128 ? --> ? 5. Why the limit vary depneding on device model ? --> ? 6. Are there any "upper limits" for ESCON or FICON ? --> ? -- snip-- John -- -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Vary devices online and offline
-- snip -- ROUTE *ALL,V ,ONLINE or OFFLINE Command processing got so backed up that automation failed to respond to some message traps for over an hour! The problem turned out to be 'aggregation', one system's attempt to gather responses from multiple systems and display them all at once. It turns out that aggregation can lead to a huge backlog of responses from ROUTEd commands even when the commands are spaced out. Workaround is a little noticed parameter T=nnn on the ROUTE command. From System Commands: -- snip -- Skip, 'little noticed parameter' seems to apply to me! I've had this problem in the past, and stopped using *ALL as the Sysplex target of the command. Instead, I simply directed the command individually to each Sysplex members (it was generated from a program - so all information was available). This got around any kind of delay issue. I'll have to look at T=00. Thanks John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Global mirror for z/OS
-- snip -- Having said that, I heartily endorse the technology. We started mirroring DASD in the late 90's when we still had to use conventional channel extenders. However you do it, nothing improves performance more than money does. ;-) -- snip -- I agree Skip. Knowing that one has a copy of their data (be it Asynch or Synch) at another site is certainly a good feeling. Distance (and cost) often dictates which method of duplication is used. -- snip -- We are building a disaster recovery datacenter about 4 hrs. from our primary site. Does anyone have any information on XRC Global mirroring for z/OS and are there any good redbooks/manuals? -- snip -- Sharon, there are a lot of good publications on the IBM site (you didn't say what kind of hardware you're using). For example. DS8000 Copy services - SG24-6787 Implementing ESS copy services - SG24-5680 (a little old - but still good) DFSS Advanced copy services - SC35-0428 Also take a look at the XRC monitor. Take a look at some of the presentations that have been offered by Share in the past (you can thank Skip for one of those). (www.share.org) Attent Share. There are often very good presentations that give you a much deeper look into the hardware than you can get from a manual. You can also hear some real user experiences. John. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: What part of z/OS is the OS?
-- snip -- So which part of Windows is the OS? File handling I/O GUI -- snip -- The mOuSe. :-> John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Correlating CHPID utilization with DASD device activity
-- snip -- For one site, the RMF Channel Path Activity report is showing substantially higher channel activity than for the other site. The times of the high channel activity are well aligned with a testing schedule, but I would like to further narrow down what is causing the extra activity. . . . Is it possible to do what I am trying to do? Is there something better to look at than the RMF DASD Activity Report? If anyone has had to do anything like this before, any ideas would be appreciated. -- snip -- Laura, RMF offers you various kinds of reports (as you have already seen). Monitor I CHAN, IOQ, DEVICE(DASD) and CACHE(SUBSYS) will give you a lot of information at the device level and you'll be able to narrow down which volumes have high usage. To drill down to the Data Set level you'll either need to startup GTF and do an I/O trace, or use one of the various Monitors (TMON, BMC, ...). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Correlating CHPID utilization with DASD device activity
-- snip -- >To drill down to the Data Set level you'll either >need to startup GTF and do an I/O trace, or use one of the various Monitors >(TMON, BMC, ...) Assuming one does a GTF I/O trace, what does one then do with the trace? Are there tools available to drill down to the Data Set level given only a GTF trace? -- snip -- In the bad old days, GPAR was available. I don't know of any other tools. I use IPCS. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Wealth of Information
There is a wealth of Information available on the Internet, but sometimes it's well hidden. Here is a collection that is very interesting reading for any DFP bigots. DFSMS Support Flashes and Technical notes. http://www.ibm.com/servers/storage/support/software/dfsms/ For example. Take a look at Technical notes point 40 - DFSMShsm Diagnostics Education. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Correlating CHPID utilization with DASD device activity
-- snip -- >>Assuming one does a GTF I/O trace, what does one then do with the trace? >>Are there tools available to drill down to the Data Set level given only a GTF trace? >In the bad old days, GPAR was available. I don't know of any other tools. I use IPCS. would, etc. I would never dream of using GTF to do data set level analysis unless I first developed, or had access to, a tool to reduce the data down to some kind of summary form and also to scan VTOCs to convert CCHHs into data set names. Both of these tasks are non-trivial. -- snip -- Yes Bill I agree, there is a huge amount of data and additional steps are needed to get meaningful results from it. If you needed to get Data Set level access information and had no OEM software available, how would you proceed? John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: DASD Performance
-- snip -- If you wanted to setup (from scratch) a DASD performance monitor (for RAID) and you weren't allowed anything but RMF and MXGSAS, what SMF records would you keep and why? -- snip -- MVS System Management Facilities (SA22-7630) is a good place to start. I would choose not only the RAID relevant records, but also Channel, Ficon Link, .. records in order to see the whole picture. Here is my short list. SMF 73 Channel path activity SMF 74 - 5 Cache Subsystem - 7 Ficon director - 8 ESS link status SMF 78 - 3 I/O queuing activity SMF 79 - Monitor II records (in case you want this kind of a snap shot). John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: RULE OF THUMB ?
-- snip -- > > If I define the vtoc (below) as > > VTOC(0,1,29) - > INDEX(2,0,45) > > The VTOC starts at cylinder 0 for 29 tracks > The INDEX starts at cylinder 2 for 45 tracks. > > If for some reason the INDEX was set to 5,0,45. Would we have a problem with response time? Is there a set rule that the INDEX must be placed after the VTOC? First, the VTOCIX in your example is much too large. It only needs to be a fraction of the size of the VTOC itself, since it is just an index into the real VTOC. The ICKDSF manual has some guidelines for sizing the VTOC/VTOCIX, but there is no penalty if the VTOCIX is too large, so with a 2 cyl VTOC, I would make the VTOCIX 1 cylinder INDEX(2,0,15). -- snip -- One of the advantages of oversizing the VTOC/VTOCIX is that when you use a fine product like FDRPAS to move devices to a larger DASD geometry (eg. 3390-9 to 3390-54), you don't have to worry about enlarging the VTOC/VTOCIX. I'm not advocating making them extreemly large, but it was very convienient for our last migration. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ICKDSF INSTALL -- What's the point?
-- snip -- My question is, "So INSTALL writes the home address and record 0. So what?" What is the benefit of doing INSTALL and INIT vs just the INIT? (All of our DASD is on virtual arrays, in case that makes a difference.) I have never bothered with the INSTALL before; is there a reason I should? -- snip -- It may be useul to use INSTALL if you planned to establish PPRC with those devices as the primary. PPRC will only send R0 in that case (8 bytes) rather than the whole track. -- snip -- PPRC is the reason our ICKDSF INIT procedure has been changed to do a medial init It was required that the R0 be properly initialised. DASD was IBM ESS. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: How many?????? - 33 PAVS to one device
-- snip -- Only 1 other device showed delays approaching this and it's on a different ssid. No enq's, no contention, no IOSQ, no high activity rates (see 1st post), no nothing. Muy Abnormal? . . That's what I was wondering too. If XRC was the real problem, adding PAV's would not help at all, but WLM probably cannot know that. However, from WLM's point of view, the device must have had IOSQ delays, otherwise WLM would not add more PAV's, but if this is the only job on the device, this is not likely. *OR* you have a WLM bug at hand. What z/OS release are you? Do you have dynamic WLM managed PAVs? -- snip -- I would concentrate on the poor I/O response time and assume that WLM is just trying to help. Are you sure that XRC isn't setting the device to long busy? If you want to elimate XRC as a possible culprit just set the device to DONOTBLOCK and test again. I would also take a look at the RMF MON I reports. BTW. There have been WLM PTFs for unnecessary PAV movement. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: XRC Primary Volume SWAP w/TDMF or FDRPAS
-- snip -- For Example; if a SWAP Source volume is also an XRC Primary volume, will TDMF/FDRPAS copy/move the volume? Or; if a SWAP Target volume is also an XRC Primary volume, will TDMF/FDRPAS copy/move the Source volume? . . Has anyone else addressed this issue? Anyone using FDRPAS/TDMF to move an XRC Primary volume? -- snip -- There is a piece in their (FDRPAS) documentation about Duplex volumes. It describes some of the limitations and recommendations. I have no personal experience with FDRPAS and XRC, but I suspect that it wouldn't be a problem. There were issues with the Data mover software not recognising a VOLSER change on the primary volume, but that has been fixed. One good reason to stop the mirroring before you move the volume with FDRPAS (probably also applies to TDMF) is that it takes longer to move the volume if it is currently in DUPLEX. The down side is that you'll have to re-synchronise the primary and duplex volumes and not have a valid copy of your data in your DR site until it's finished. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Antwort: Re: XRC Primary Volume SWAP w/TDMF or FDRPAS
-- snip -- As I said, we modify the label on the source volume so that it can't be brought online, so IPL will ignore it and only see the volume on the target disk. > And I wonder what XRC sees during the SWAP with the Target offline, > but since XRC is controller based it should see all of the I/O traffic from > FDRPAS. If you have established a XRC session on the FDRPAS target disk prior to the swap, then all the WRITEs to that disk should be mirrored,so the XRC secondary will be an accurate mirror of the FDRPAS target disk when it is brought online. FDRPAS does no WRITEs to the source disk except for modifying the volume label at the end, so that change should also be mirrored by that XRC session. -- snip -- One potential problem that I see is that the XRC system (after the swap has happened) thinks that it has two primary volumes that both have the same volser. XRC doesn't propagate the I/O's that modified the volume label to the secondary volume, but instead updates its own duplex pair information. This information is then used during XRECOVER processing to relabel the secondary volume to the correct primary volser. Does XRC see the label change in the primary system as a normal ICKDSF reformat? If yes, then the duplex pair information will also be updated (at a minimum you'll get an ANT message on the secondary system telling you that the volser has changed). If no (I don't know what FDRPAS actually changes), then XRC will continue to believe that this primary/secondary pair are associated with the original volser. That would cause problems during XRECOVER processing. I don't have an XRC system to test this anymore, but Donn, you could easily test this yourself and let us know what exactly happens. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: PPRC status display & reporting
-- snip -- I would like to get documentation about the info returned by the ANTRQST macro. I am using the "ANTRQST ILK=PPRC,REQUEST=PQUERY,..." macro instruction, and have tried both the FORMAT=LONG and the FORMAT=PQMAP parameter. In both cases I have had to guess the layout and/or the meaning of the returned information. My source of information so far has been the "z/OS DFSMS Advanced Copy Services" manual. We are running z/OS 1.6. Is there any documentation with detailed information about the info returned by ANTRQST? -- snip -- The unformatted output is mapped by the ANTPQMAP macro in sys1.maclib. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: FICON Saturation?
-- snip -- I'm seeing a DFSORT that normally runs in 6 minutes take 33 minutes with the same record counts. It's accessing 1 disk that has an extremely low activity rate, response times not that high, IOSQ time appears to be 95% of resp time and device delays of 99 - 100%. One VSAM cluster is being -- snip -- A high IOSQ doesn't sound good. Does this device have any PAVs assigned to it (look at RMF)? Are there any other tasks that are also using this volume? Is it possibly a duplexed volume in XRC that is being put on long busy due to high IO activity in general (ANT... messages in SYSLOG)? RMF will also show you FICON usage if you want to eliminate that as a possible cause. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: HSM Missing Member from Recalled Dataset
--- snip --- When he recalled it again, the member that he had added was missing. I checked through SMF and saw the activity. The fact that the dataset ([data.set.name]) was on [volser] (a NONSMS volume) complicates matters a little. --- snip --- What kind of activity did you see in SMF that shows that a data set had an additional member added! If HSM reconnected the ML2 version, then I would expect that the original data set was not changed. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Help w/HSM Autobackup and Concurrent Copy
-- snip -- I currently have concurrent copy preferred in all of my management classes. I noticed during autobackups that datasets in use receive a RC of 8 from concurrent copy due to, I'm assuming, the TOL(ENQF) parameter not being included in the defaults for the dump command. HSM then issues an HBACK for the dataset and it gets put on an ML1 volume. Is there a way to change the default values HSM uses for ADRDSSU and if there is should I change them What do others do? Thanks for the help. -- snip -- Take a look at DFSMShsm Storage Administration Guide chapter 'Specifying How DFSMShsm Should Respond to a Data Set in Use'. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Z/Os Storage Mgmt products
-- snip -- > Responses mentioned the following products: > Tivoli Omegamon XE for Storage, Tivoli Storage Optimizer, CA-Vantage > Storage Resource Manager, BETA85, DTS, Mainstar products... > > It raises a question: what are the needs ? > IMHO those products are quite different, they address different issues. > > Last but not least: is ANY product really needed ? Right!, that was my first reaction too. What are your problems and do you need to solve them with costly softwaretools and manpower or could it be more costefficient to add a couple of TB's to the SMS pool(s) and let SMS manage the whole lot? -- snip -- If all you want to manage is x37 Abends, then I agree. But these tools do much more than that. Let me give you an example of what CA-Vantage can do (I'm NOT on the CA payroll by the way!) Pool monitoring (define your own filters and break down the usage for each department - DASD or TAPE). Trend this data (eg. look at DASD growth over the last past years - broken down into which ever categories you want). Notification when certain events occur (eg. GDPS Primary volume no longer in PPRC Duplex, or HSM CDS backup took longer than xx minutes to complete). Take what ever action you deem appropriate (simple Email, or MVS commands, ..) Examine HSM migration and report on thrashing between lvl0 and ML0 Automatically build (and submit) jobs to correct the management classes of certain categories of data sets. the list goes on. A very powerful product. The only tool I've worked with is CA-Vantage, so I can't speak for any of the others, but I suspect that they all have their selling points. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: z/OS 1.8 Features
-- snip -- Since most folks don't get a chance to explore the new features, thought I'd list the features introduced over the last couple of z/OS releases that I think will be of interest to applications programmers. -- much good stuff snipped -- Very well done Steve and much appreciated (even the non-application types will find this useful). Thanks John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ESS pprc question
-- snip -- > We have two ESS800 in two locations with PPRC v2. We plan to power-off > ESS_A (primary) for realocation (1-2 hours). For this period of time > system and application will be stopped. > During this time we don’t plan to start system in secondary location, so > there will be no update on ESS_B > I’m looking for a safe procedure to power-off ESS_A with existing PPRC > and I don’t want to perform ‘initial copy’ after procedure. My opinion, assuming all systems are down during the relocation: Method 1: - Stopping all systems leaves the volumes in sync. You don't have to do anything. -- snip -- That has also been my experience. Just power them down normally. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ACS Filtlist overhead
-- snip -- Even with a large FILTLIST, I don't think that the overhead needs to be worried about. The ACS routines are driven mainly during dataset allocation, so unless your site allocates an extraordinary number of datasets per day, the overhead is probably not noticable. Compared to the rest of the code for allocation (volume selection, DADSM, catalog, etc.) I doubt if even a large filtlist will be a significant part. You could put tests up front in the ACS rtns to identify datasets that definitely don't match the filtlists and bypass that search for many datasets. -- snip -- A very different approach is to use RACF to contain information that you can then check in your ACS routines. It may be a better approach, since you will not have to maintain very large filter lists. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: HSM control dataset control block map's
-- snip -- If someone knows where can I find a map of the DFHSM Control Data Set records (MCD etc ) -- snip -- Try DFSMShsm Diagnosis Reference Document Number LY27-9608 John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: HSM migrate
-- snip -- I could manually migrate the dataset off but what is stopping HSM doing it automatically? Any ideas? Thanking in advance. -- snip -- Take a look at your SMS Storage Group Migrate High/Low settings. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Allocated volumes after the power loss in the storage subsyst em
-- snip -- There was a UPS problem, it registered(?) anomaly in the electricity and itself and cut the electricity on the site (that was my understanding). I came later in and had to do only with the problems on the mainframe side. The simplex devices (that is, not mirrored) were not critical for the production (the applications could go on with the half of the capacity). The only problem was that applications hanged as result of these allocations. Luckily, it happened in the night, and as the electricity came, we had the access to all the volumes again - no problems. Well, as Ed Finnell wrote: "when the subsystem gets IML'd (takes forever under normal circumstances)". Normally, it is a primary GDPS site for one production sysplex, but at the time that was our secondary site, so we lost only the systems running there. -- snip -- I've lost track here! You're running GDPS (I assume with PPRC and Hyperswap), and you lost power to your secondary site. I assume that those devices were also the secondary Duplex devices. Is it possible that you were experiencing a Long Busy situation in your production site? (admittedly - very long!) John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
3390-81
Hello List, I've come across a couple of references to a 3390-81. Now, we have -27 and -54 devices. Have I missed some announcement? John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Replacement for 3179 terminals locally attached to a 3174 controller
-- snip -- Don't define NIP consoles. Let it fall through to the HMC. Works really well and you always know where the NIP console is going to be. >> What if the HMC fails? -- snip -- You always have a service element that can accept NIP messages. Depending on how you set up a GDPS environment, it can be a requirement to ensure that no NIP consoles are active. John. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Looking for documents
-- snip -- See: SG24-6898-00. It's about Logger, but Chapter 9 mentions those two manuals about GDPS: GDPS/PPRC V2R8 Installation and Customization Guide, ZG24-6703 GDPS/XRC V2R8 Installation and Customization Guide, ZG24-6704 Since the books are mentioned in Redbook, I assumed they're available to public. -- snip -- Top secret. As far as I know, they are only available to customers that are implementing GDPS and have agreed to not disclose any proprietary information. ZG is correct. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: MOD54
-- snip -- Any one with 3390 mod54 . Any problems experiences good or bad -- snip -- No software or response time problems to date. Ensure that you check with your Software suppliers for support info. Dumping them takes a very long time. We're reviewing whether it is needed at all in our particular case. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Restarting SMSVSAM address space
-- snip -- Hello, I am setting up environment for RLS. I started SMSVSAM (with V SMS,SMSVSAM,ACTIVE), but now I have to restart it (had some security problems), and I don't know how. I issued V SMS,SMSVSAM,ACTIVE again, since the explanation of the command states that 'it restarts SMSVSAM address space', so I expected it to shut it down and start again. There was no response to this command. Then I tried V SMS,SMSVSAM,TERMINATESERVER, got the message *IGW572I REQUEST TO TERMINATE SMSVSAM ADDRESS SPACE IS ACCEPTED: SMSVSAM SERVER TERMINATION SCHEDULED. Nothing happened after that command, SMSVSAM is still active. -- snip -- If TERMINATESERVER didn't work, you may need to issue FORCE,ARM or FORCE (at your own risk of course). Are there any outstanding messages? Are you up-to-date on maintenance? I got hit by some SMSVSAM issues in termination that have since been fixed by IBM. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Applicable component levels
-- snip -- I've just read an APAR OA11094 description. At the end there's a list of PTFs for various levels of RMM. R1GB PSY UA19542 UP05/08/02 P F508 R1GC PSY UA19543 UP05/08/02 P F508 R1G0 PSY UA19532 UP05/08/02 P F508 R1HB PSY UA19557 UP05/08/02 P F508 R1HC PSY UA19558 UP05/08/02 P F508 R1H0 PSY UA19555 UP05/08/02 P F508 R1JB PSY UA19560 UP05/08/02 P F508 R1JC PSY UA19579 UP05/08/02 P F508 R1J0 PSY UA19559 UP05/08/02 P F508 R1KB PSY UA19594 UP05/08/02 P F508 R1KC PSY UA19595 UP05/08/02 P F508 R1K0 PSY UA19593 UP05/08/02 P F508 How can I find out what PTF should be used for z/OS V1R4 ??? What does is recorded in first column ??? It's not system version, it is not RMM FMID. What is PSY, what is F508 ? Why it is so incomprehensible ? -- snip -- For z/OS 1.6 FMID HDZ11J0 corresponds to R1J0 FMID JDZ11JB corresponds to R1JB You would need R1G0 for the base function. I agree - it's a PITA. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: JES2 proclibs allocated???
-- snip -- > 1) TDMF process have moved lib03 to another DASD. > 2) This lib03 has been deleted (manually). > 3) This lib03 has been allocated (through 3.2 option) from system "x" > > Then, some messages are showed: > IEC143I 213-04,,JES2,JES2,PROC00-0003,,volser > > How could fix it? AFAIK, JES2 restart is required. -- snip -- A Batch Job with /*JOBPARM PROCLIB=... for each intrdr should also work. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: manipulating catalogs
> > The Managing Catalogs manual describes this sequence with a DELETE and a > DEFINE between the export and import, so the import is into an empty > catalog. In this case, jobs referencing the catalog would probably get > upset. > > Why do you wnat to do this if you are not going to empty the catalog > first? > > Jim McAlpine <[EMAIL PROTECTED]> wrote: > Is it ok to do the following - > > LOCK CATALOG > EXPORT CATALOG > IMPORT CATALOG > UNLOCK CATALOG > > while there are jobs referencing datasets in the said catalog, or should I > wait until the work is quiesced. > > Jim McAlpine > Many years ago, a customer was following this procedure. Soon after the UNLOCK CATALOG, jobs started have catalog problems. It was deemed that the Ucat was corrupt, and a backup copy was restored (using LOCK/UNLOCK). The problems persisted. OPC (Job scheduling tool) was defined as Privileged in the ICHRIN03, so the Catalog locking wasn't observed. It took a while to clean up. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Trying to catch output from HSEND command
-- snip -- > But i find that the OUTTPUT function is working only with PUTLINE macro > no with TPUT macro witch is used by HSM for generate mess. > Can somebody tell witch function to use for catch HSM messeges generate > by TPUT macro? I'm not aware of any standard function you can use from a batch TSO job that will trap output generated by TPUT. Perhaps, as an alternative, your exec could establish a CONSOLE environment, then perform your HSM function using operator commands, and retrieve the command output from the console. -- snip -- I think we've had this question before. One method I use it to direct the HSM output to a data set (use ODS on the command) and then set up a loop to wait until the exclusive ENQ has been released by HSM. (Use address syscall "sleep 1" and check x = SYSDSN(hsm_out) for 'OK') Not pretty, but it works. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: Trying to catch output from HSEND command
-- snip -- I raised a share request for a supported API into HSM information many years ago - but I believe this died a death due to lack of interest from anyone but me. -- snip -- Rob, and you'd problably get the same kind of API implementation that the System Data Mover (ANTRQST) got. Namely, the buffer information returned for a query request is exactly the same as the TSO output generated via a command. You get to parse it anyway!! I just took a look at the z/OS 1.8 documentation and there is now an RQUERY request and mapping macros for the output. I'm glad it's been improved. I also second your HSM API request. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: RESERVE and ENQ (GRS question)
-- snip -- >I believe ENQ is not needed, but I'm not sure. A reserve only happens if there is an ENQ. It's one bit in the control block. You cannot have a reserve by itself, I believe. -- snip -- I remember being surprised when I coded a RESERVE and nothing happened. Allocating and reading the VTOC generated the desired reserve. I don't think that it was the ENQ but the I/O to the device that caused the reserve to happen. John. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: RESERVE and ENQ (GRS question)
-- snip -- > Correct, the RESERVE macro only sets a bit that the next CCW chain to > the device should include the CCW to effectuate the Reserve. Doesn't it depend on SYCHNRES setting in GRS ? BTW: I think it is unrelated to my doubt. My questions is (I'll try to rephrase it) : Does RESERVE cause lock on local system as regular ENQ does ? -- snip -- I forgot about SYNCHRES. Yes you are right it depends on the setting. I'm not sure what you mean by 'cause lock on local system'. When you code the RESERVE macro, you specify the qname and rname. In addition to the Hardware reserve, you will also get an enqueue for the qname/rname combination. If you add the SYSTEMS parameter you'll get a global enqueue. For example.RESERVE (QNAME,VOLUME,E,6,SYSTEMS),UCB=UCB_PTR,LOC=ANY I actually just tried the macro with no SYSTEMS parm. but the expansion is the same. So I suspect that the enqueue is always a global enqueue. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ISGAMF00 problem
-- snip -- I'm trying to trace some SCOPE=STEP ENQ's. I RTFMed. I started GRS Monitor (ISGRUNAU). I modified ISGAMF00 member by putting GFLG FILTER=N. Then assembled, linkedited, issued F LLA,REFRESH, then F GRSMON,I=02 (my sufix). However I can't see any ENQ with scope=STEP. I'm pretty sure there should be any, because I submitted some job issuing such ENQs. -- snip -- I haven't used the monitor for a very long time. It seems to have grown up a little! Firstly, confirm that these ENQs do exist (D GRS command, or take an SVCDUMP and peek at the GRS control blocks). Try setting up a filter with GFLAG .. ,STEP=Y to let the scope STEP resources through. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: ISGAMF00 problem
-- snip -- > > Although the documentation and parameters would seem to indicate > otherwise, > the GRS monitor is not capable of recording STEP ENQs. Scott, Thnk you for the explanation. Indeed - I'm not happy of such surprise. However I have on more question: Can I see scope=STEP enq's using D GRS,RES=(q,r) command ? My experiments say I cannot. Or, more general: is there any method to dsiplay STEP enqueues ? -- snip-- That suprises me. GRSDATA via IPCS in an SVCDUMP definetly shows you the STEP ENQs. (from the IPCS manual) Resources are presented in the following order: 1. ASID(X'') (STEP) resources (ordered by ASID) 2. Local (SYSTEM) resources 3. Global (SYSTEMS) resources This is consistent with the order used by verb exit QCBTRACE in prior releases and with the order used by the GRSDATA subcommand in the current release when GRS control blocks are used instead of the data collected with the SDATA=GRSQ option of SDUMP. -- snip-- > > Choose your favorite flame mechanism. I'd suggest a RCF. I think Shane just chose his. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: BLKSIZE=0
--- snip -- Bruce, This conflicts with BLKSIZE=0 being used for SDB. SDB does not work unless DSORG can be established. BLKSIZE=0 is a null - SDB requires DSORG and no BLKSIZE. Additional DCB depends on the DSORG. Most sites have a default DATACLAS that will allocate with DSORG=PS if nothing else can be determined. This allows SDB to work, and also ensures SMS datasets have an EOF at allocation. Ron > (the DSORG is not > determined until the dataset is opened for output -- snip -- We've just noticed (z/OS 1.8) that a problem was fixed with DSORG=PO data sets and dynamic multivolume. Our ACS routines (as probably other sites also do) check for DSORG. A DATACLAS with dynamic multivolume capability is assigned to most data sets, unless they don't support mulitvolume (DSORG=PO). This had worked well in the past. Data sets that have an unknown DSORG also had a DATACLAS assigned to them (with DYNVOL) specified. Unfortunately, some of the unknown DSORG data sets happened to be PDS data sets that were allocated with the LIKE parameter (DSORG is NOT passed to the ACS routine when LIKE is used). This now results in an allocation error. Be careful with assigning a DATACLAS to an unknown DSORG allocation in SMS. John PS. Good to hear that you're still around Ron. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: z/OS 1.7 SYNCHRES=YES & z/OS 14 SYNCHRES=NO
-- snip -- I have checked the manuals and the archives on this and I am still unsure about mixing the yes and no values for SYNCHRES in a GRSPlex (Basic sysplex if that matters any). I need to bring my TEST LPAR into . . -- snip -- We migrated from z/OS R4 to z/OS R6 and allowed the default to be in effect so we implemented SYNCRES=YES on each system at the same time as z/OS R6. We ran a mixed level Sysplex for many months as the migration of R6 onto a couple production LPARs had some interesting technical challenges. . .We saw no ill effects or need to override the default in R6 till we implemented it across the Sysplex. -- snip -- I can't speak for this particular option, but in general there is great peril in running shared systems with different views of what to protect/allow and how to do it. In some cases, a new system cannot even join a GRSplex if key options don't match. You could take the trusting optimist view that if joining is not prohibited, then everything must be hunky dory. -- snip -- RESERVE processing is different with SYNCHRES and I had problems early on (with XRC and CU long busy - if I remember correctly). There is no requirement for SYNCHRES to be the same in all members of the Sysplex. But, as Skip points out, it is preferable if all members of the Sysplex serialize their resources in the same manner. I would upgrade your z/OS 1.4 systems to using SYNCHRES=YES before running the z/OS 1.7 System. As Sam indicated, it also works in a mixed Sysplex, but I would prefer to migrate to SYNCHRES=YES first, before upgrading the z/OS image to 1.7. It would be nice to know if you have any RESERVE/SYNCHRES problems before migrating. For additional reading, look at OA18968 (z/OS 1.8) and also OA14080. The last apar is especially interesting since it deals with LSPACE, MIH and SYNCHRES=YES. John -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
