Re: Allocation of an SMS managed dataset
Lennie, VTOC / Index will be first. Anything with VVDS will come later. Thanks. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DS8K features
"Read the Data Sheet" on https://www.ibm.com/products/ds8000#:~:text=IBM%20Storage%20DS8000%20is%20the%20fastest,%20most documents that compression is "up to 2:1". DS8K does not dedup. The use case for DS8K onchip compression is for data not compressed (zEDC) nor encrypted on z/OS. Since the storage is thinly provisioned, DFSMS will provide support to help clients manage the difference between logical and physical storage consumption. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Allocation of an SMS managed dataset
Please note that it is on our 'short list' of improvements to avoid selecting volumes that will fail data set allocation due to a full VTOC or VTOC Index. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Anyone exploiting ZEDC?
DFSMShsm is an excellent use case for zEDC and is our number one best practice for HSM. When enabled, DSS zEDC compresses all blocks of data passed to HSM during Migration and Backup. Because HSM is processing fewer data blocks, both cpu and elapsed time are reduced. When going to ML1, the amount of storage is also significantly reduced. Glenn Wilcock, DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: IBM VTS and cloud
Another option is IBM Cloud Tape Connector. One of its capabilities is Virtual Tape Replication, which replicates tape data sets to cloud object storage. If interested in finding out more, please email me at wilc...@us.ibm.com and I can connect you with the SME. Thanks. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: IBM VTS and cloud
I sent your question to Joe Swingler, TS7700 architect and replied as follows: "As of today, a single TS7700 can be either cloud attached or tape attached, but not both. So, a third TS7700C would be needed or one of the two tape attach boxes would have to become only cloud attached." Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: What happens in HSM when I change a Management Class
I should have added that EXPIREBV will run for a while if you haven't run it for some time. So, do it over a weekend. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: What happens in HSM when I change a Management Class
You'll need to run EXPIREBV. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMSHSM reports.
Thanks all for the great comments. A few additional comments: 1) One of the sample reports for the DFSMS Report Generator is a Migrate/Recall thrashing report. 2) By default, HSM does not create an FSR record for data sets that are not eligible for primary space management. These also cause 'spin' because they are examined every cycle but never processed. Create FSRs with the following patch from the DFSMShsm I&C guide: "Enabling FSR records to be recorded for errors, reported by message ARC0734I, found during SMS data set eligibility checking for primary space management" To specify that FSR records are to be recorded for errors, reported by message ARC0734I, found during SMS data set eligibility checking for primary space management, issue the following PATCH command: PATCH .MGCB.+EF BITS(..1.) To specify that the FSR records are not to be recorded, issue the following PATCH command: PATCH .MGCB.+EF BITS(..0.) 3) Also documented in the DFSMS I&C, a patch to enable HSM to process empty and certain undefined data sets "Enabling backup and migration/expiration processing of empty data sets" DFSMShsm can be tuned such that the storage administrator can request that certain data sets that are allocated but not opened be backed up, migrated, and/or expired based on typical DFSMShsm configurations. The following two data set variations that can be handled by this change: 1. Cataloged data sets that have been allocated such that DSORG=0 (undefined) and have not been opened for output (DS1DSCHA=OFF and DS1LSTAR=0). This support will allow these types of data sets to be backed up and migrated based on the local DFSMShsm configuration. 2. Data sets that have been allocated such that a nonzero DSORG is set, but the data sets have not been opened for output (DS1DSCHA=OFF and DS1LSTAR=0). This data set variation is typically supported by DFSMShsm but might not be backed up as the change indicator (DS1DSCHA) indicates that the data set is unchanged. When DS1DSCHA indicates that the data set is unchanged, it can prevent the data set from being expired if the environment requires a valid backup before expiration. Once enabled, this support can ensure that these types of data sets can be backed up before they are migrated, thus allowing them to be expired based on the local SMS and/or non-SMS configuration. To support the data set variations that were just discussed, issue the following patch commands: PATCH .MCVT.+287 BITS(..1.) Request empty non-VSAM data sets (DS1DSCHA=OFF & DS1LSTAR=0) to be backed up. PATCH .MCVT.+287 BITS(...1) Allow empty cataloged data sets (DS1DSCHA=OFF & DS1LSTAR=0) with DSORG=UNDEFINED to be backed up and migrated Note: The following considerations must be made to ensure minimal impact to your environment as well as desired results: In environments that have many of these newly supported data sets, DFSMShsm functions can require additional time and CPU to complete. If these data sets have accumulated incrementally over time, this additional time and CPU requirement might be temporary. If these data sets are created frequently and in large quantities, the impact to DFSMShsm functions might be permanent. In a SETSYS INCREMENTALBACKUP(CHANGEDONLY) environment, DFSMShsm relies upon the DS1DSCHA flag of a data set to determine when the data set requires backup and when a backup might have already been created. With this support, empty data sets can be backed up even when the DS1DSCHA flag is off. For this data set type, DFSMShsm will rely on the backup date of the backup version to determine whether the data set with DS1DSCHA has a valid backup. For data sets that are allocated, backed up, deleted, and reallocated on the same date, DFSMShsm will not be able to distinguish that the backup is from a previously created version of the data set. The backup made on the same day as the creation will qualify the data set for expiration if the MC requires a backup or if a MIGRATE DELETEIFBACKEDUP command targets the data set. Given this support only applies to empty data set variations, this issue would only impact data sets whose most recent version is empty. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DSS dump and migrated datasets
Thanks Mike. ABARS is great for this type of use case. ABACKUP with the MOVE option will backup all data types (no recall for migrated data) and delete the original copies, including the migrated data sets. Thus creating a single backup of all of the related data. Essentially, an Archive type of function. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Cloud Storage
There is also DS8000 Transparent Cloud Tiering. For TS7700 Clients, TS7700 Cloud Storage Tier is an excellent way to maintain SLAs and also leverage cloud storage for cold data. Overview of IBM z/OS object storage solutions: https://community.ibm.com/community/user/ibmz-and-linuxone/viewdocument/zos-cloud-object-storage-solutions?CommunityKey=3c6c69c1-abde-4ee6-b603-19ef14942e8f&tab=librarydocuments -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Z skills training
Besides formal training, also leverage the free z/OS Academy Events: "https://community.ibm.com/community/user/ibmz-and-linuxone/groups/community-home?communitykey=4f04f510-4165-4fa6-8a82-c00b4d150967";, join SHARE and gain access to its library of recordings: "https://www.share.org/Education";, and specific to DFSMS, there is a library of educational recordings: "https://ibm.ent.box.com/s/w2jpq95ydynpztbjwdplw3l0hoipl0i8";, "https://ibm.ent.box.com/v/DFSMS-Academ-DFSMShsm2021"; and the DFSMS Community: "https://community.ibm.com/community/user/ibmz-and-linuxone/groups/topic-home/blog-entries?communitykey=3c6c69c1-abde-4ee6-b603-19ef14942e8f";. Glenn Wilcock, DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Permanent Incremental FlashCopy Relationships? (Was: SETROPTS ERASE...)
Hey Ed, we haven't really been advertising this function because of the DS8K copy services restriction. Besides FlashCopy, it means that only simplex volumes are eligible (no Metro or Global Mirror relationships). When creating FlashCopy Full Volume copies, Incremental FlashCopy is still a best practice. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMS Advanced Customization Guide
Hi Bill, this is a licensed manual. Please direct questions to its owner, Andrew Wilt. anw...@us.ibm.com Thx. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: z/OS Client Web Enablement Toolkit
For those interested in leveraging cloud object storage, please refer to the z/OS 3.1 preview announce that included a reference to a new DFSMS access method for cloud data - Cloud Data Access (CDA). It is built on top of the Web Enablement Toolkit and simplifies authentication and put/get/etc operations to the various cloud object storage providers. It will be provided at no cost. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Are Banks Breaking Up With Mainframes? | Forbes
Redpaper: https://www.redbooks.ibm.com/redpapers/pdfs/redp5705.pdf Abstract: In the recent Application modernization on the mainframe IBM Institute for Business Value study, among the surveyed, 71% of executives say that mainframe-based applications are central to their business strategy. As stated earlier, four out of five respondents say that their organizations must rapidly transform to keep up with competition, which includes modernizing mainframe-based apps and adopting an open approach that includes cloud. IBM can help you meet these challenges by providing the tools and methodologies to help you transform your business. To increase this productivity, agility, and close the skills gaps, you should modernize your mainframe applications and embrace hybrid cloud. Specifically, consider the following best practices: Embracing a hybrid cloud approach to mainframe application modernization. Leveraging a continuous application modernization journey. Taking advantage of the recommended cloud reference architectures to leverage IBM Cloud®, Amazon Web Services (AWS), and Microsoft Azure. Getting started fast with the IBM proven co-creation methodology -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Are Banks Breaking Up With Mainframes? | Forbes
This post from Ross is a must read for this topic: https://www.linkedin.com/pulse/results-mainframe-application-modernization-migration-ross-mauri/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Command/interface to find the total size of a MIGRAT2 disk file?
For future reference, HSM DCOLLECT reports are also available if you want reports for larger amounts of information. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Tape compression and modern encryption facilities.
This discussion is all about use cases and data types. As mentioned in this thread, not all data supports zEDC and it's highly unlikely that everything will be encrypted. I would recommend analyzing the data written to your tape environment and determine if it would be beneficial to turn off tape compression for a specific workload. (Via data class). DSS was one use case that was mentioned. I would never use the COMPRESS option of DSS if you have the zEDC feature. If you are targeting disk, then use ZCOMPRESS (zEDC). If you are targeting tape, then it's a trade-off with using zEDC and TS7700 compression. zEDC will use host cpu. DSS has the smarts not to try to compress a data set that is already compressed and or encrypted. If cpu consumption is a concern, then just use TS7700 compression and no DSS compression. This is especially true for full volume dump to tape. HSM is a use case for which zEDC is a best practice. It will improve throughput and slightly reduce CPU. We recommend that you use ZCOMPRESS for both migration and backup. (If you use HSM Dump, then it's the same recommendation as DSS. Just use TS7700 compression to not experience a cpu increase). For your HSM migration/backup data classes, disable TS7700 compression. For Recycle, assign a data class that utilizes TS7700 compression because you need it to continue to compress old data that wasn't migrated/backed up with zEDC. For other tape data, determine if the application is already compressing and or encrypting the data and enable/disable TS7700 compression as appropriate. Glenn Wilcock, DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Hsm for system dump volume
Yes, auto migration will suffice. Use the HSM ADDVOL command to have HSM manage a nonSMS volume. You'll define it as a PRIMARY volume. You can specify the days to keep the data sets on the volume before they are migrated with the MIGRATE(days) keyword. You should also specify the THRESHOLD command to indicate high and low thresholds. Backup only applies if you want to additionally create a backup copy of the data sets. Backup doesn't delete the data sets after the backup, so you'll essentially have 2 copies. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Transmitting SMF records
Late to the party on this discussion... DFSMS recently shipped a new utility named GDKUTIL. It converts z/OS data sets and UNIX files to cloud objects and visa-versa. This enables utilizing cloud object storage as a method for z/OS sharing data, such as SMF, as opposed to FTPing. Use GDKUTIL to write the data to object storage for shared access across sysplexes / platforms and visa-versa - ingest off-platform data into z/OS. DFSMS CDA (Cloud Data Access) is used to securely manage cloud access. https://public.dhe.ibm.com/eserver/zseries/zos/DFSMS/CDA/OA62318/Cloud_Object_Utility_Pub_updates.pdf Glenn Wilcock, DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: IBM Solutions for leveraging cloud object storage on z/OS
Hi Tom, you're not alone! I've added an attachment of the pdf at the same link. Viewing the pdf in full-screen mode gives you the same navigation, but removes all of the animation so that you can go at your own speed... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
IBM Solutions for leveraging cloud object storage on z/OS
I often get asked what solutions are available to leverage cloud object storage solutions on z/OS. I'm working on a slideshow introduction to the solutions that IBM offers. I plan to add a voice-over and some navigational enhancements, but wanted to share what I have so far. Please feel free to reach out for follow-up discussions. (I'm taking the last 2 weeks of the year off, so there will be a delay in my response over the next 2 weeks). Happy Holidays and New Year! (The slideshow is stored on the IBM Community. When I tested the link, it didn't require a log on). https://community.ibm.com/community/user/ibmz-and-linuxone/viewdocument/zos-cloud-object-storage-solutions?CommunityKey=3c6c69c1-abde-4ee6-b603-19ef14942e8f&tab=librarydocuments -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: MCDS CDS BACKVOL fails
Hey Jake, Even if you didn't get an ARC0745, I recommend following the steps of that message. You need to see what your current backup copies are cataloged as. If you don't have any, allocate as many as backup copies that you will be keeping. Notice that you have to allocate them for the MCDS, BCDS, OCDS and JRNL. I believe that you need to start with V001. (For grins, you may start with V000 and then allocate an extra on the end) Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: File Transfer Alternatives - Different Platforms
For folks on this thread who would be interested in doing this via cloud object storage, DFSMS is looking for sponsor users. You'll need an NDA to participate. Please reply to me via wilc...@us.ibm.com if you're interested. Thx. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM backups and CICS VSAM files,
If the data set is defined as BWO eligible, then DSS logic (data mover for HSM) can handle backing the data set up while open.How frequently a backup copy is created is determined by management class (SMS) or HSM policy (nonSMS). If a CI/CA split occurs during the backup, then it will be failed. It can be automatically retried with the HSM INUSE option. (Search for 'BWO' in the HSM Storage Administration manual) and it will take you to the INUSE section. Search for 'BWO' in the DFSMSdss Storage Administration for all of the gory details :) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: How to keep the response from HSENDCMD batch
Glad that you were able to solve it! If you have other questions, please feel free to reach out directly... wilc...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: How to keep the response from HSENDCMD batch
Hi, I'm not familiar with QUERYSET, but there is a single thread that processes certain commands (QUERY, LIST, etc), so it's possible that the QUERY was queued up behind one or more other commands that took longer to run. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: How to keep the response from HSENDCMD batch
FYI - Here's a link to 8+ hours of video recordings of DFSMShsm Education: https://ibm.ent.box.com/v/DFSMS-Academ-DFSMShsm2021/folder/133296597964 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: directory backup
Yes, the access and modification times are preserved. But, the change time is updated because the metadata is changed due to update the last backup timestamp. But, as indicated, if the wildcarding selects only those files to be archived, the backup and delete can be done with a single command. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: directory backup
DFSMShsm supports wildcards and has exclude capabilities. It will even do a DELETE if you want to delete all of the files after the backup. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Resizing MCDS
Great discussion here. For those who are interested, here is a link to charts/recordings for a DFSMShsm education series. One of the sessions is specifically on the DFSMShsm control data sets that discusses RLS, CA Reclaim, Reorgs, etc. Feel free to contact me with any questions. https://ibm.ent.box.com/v/DFSMS-Academ-DFSMShsm2021 Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: zEDC Justification documents or links
Hi Robert, if your shop uses DFSMShsm, I have some charts that I can share with you that show the significant advantages of using zEDC for backup and migration. zEDC is the #1 Best Practice for DFSMShsm. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: automated SMS Storage pool management
Dave, HSM has 2 methods to dynamically address volumes going over threshold: Interval Migration and On Demand Migration. Interval Migration runs hourly and processes volumes that went over threshold during the previous hour. The preferred method is the On Demand Migration technique. With ODM, as soon as a volume goes over threshold, SMS throws an ENF that HSM listens to and immediately space manages the volume. Windows can be defined to quiesce this activity during specific times, such as when critical production workloads are running. It's time to add more storage when the HSM space management runs to completion and a volume still doesn't get under threshold. If you're new to DFSMS, we have a series of recorded DFSMS Academy sessions: https://ibm.ent.box.com/s/w2jpq95ydynpztbjwdplw3l0hoipl0i8 If you're new to DFSMShsm, we have an entire series of recorded DFSMShsm Academy sessions: https://ibm.ent.box.com/v/DFSMS-Academ-DFSMShsm2021 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: ADRDSSU and encrypted files
Hi Cameron, to answer your base question... No, ADRDSSU does not support decrypting an encrypted file during the dump process. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EAV and zHPF documentation, anyone?
There is a Redbook for EAV: https://www.redbooks.ibm.com/abstracts/redp5364.html?Open -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM dates?
HSM GA'd March 31, 1978. I don't know when release 2 went out. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Secondary sources for DFP and DFSMS
"If you don't quote at least a tiny bit of what you are saying no to it is difficult to tell what the question was, and as a result difficult to gain much insight from your answer. Charles" Charles, it was in reply to this thread... What's the magic to getting the original message included in the reply? I was assuming the "Reply" would include it. Guess not. Thx. "Would you have articles, life cycle dates or announcement letters for any of 5645-001 5647-A01 5655-068 5655-069..." -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Secondary sources for DFP and DFSMS
Unfortunately, no. As the product owner of DFSMShsm, I was passed down the original announce for the product, but it's the internal IBM Confidential version. You think that we would have all of this stuff Migrated off to a reel tape somewhere :) Friday fun fact... In an informal survey that I recently did of the clients that I regularly work with, the oldest migrated data set still managed by HSM was migrated 11 Dec 1980!! Before many of today's other platforms even existed! -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Secondary sources for DFP and DFSMS
We've been vexed with this for some time. We, the DFSMS org, have wanted to update this for some time. But, we've run into these same issues, probably even more so since we are the product owners. So, first, thanks so much for working on this! SHARE presentations are a good idea, but wondering if since those are created by the product owners, if they will pass the litmus test. I've thought about Redbooks since we sometimes get non IBMers as authors, but didn't find any offhand that weren't written by IBMers. Those familiar with Wikipedia, can a secondary resource come from someone within the IBM company, but not directly related to the referenced products? If so, then the 'ABCs of IBM Systems Program Volume 3', is a great reference. Outside of these sources, our google searches haven't provided much as far as secondary sources besides billable educational offerings. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Using HBACKDS or OMVS hbackup for USS files?
Hi, glad that you got past the problem. Would you mind sharing what solved the issue? I wouldn't expect that a setup issue to cause an 0C4 Abend, so I would like to see if there is something we can do on our end to detect the condition and have more user friendly messaging. Thanks, Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Flashcopy for Non-SMS volume backup?
Ed, Glad to hear that it is working out for you! TCT will be a perfect addition. 1) HSM does provide data set level recovery from copy pools, just differently :) If the data set has not been deleted or moved since the backup, it's supported w/o a catalog capture. To support recovery for deleted/moved data sets, then it requires the data in the copy pool to be cataloged to catalogs also within the copy pool. You then define the catalogs to the copy pool and indicate that HSM should capture the catalog information during the FRBACKUP. This will extend the elapsed time of the FRBACKUP. LIST COPYPOOL has a datasets option to list the data sets that were included in the FRBACKUP. You then use the DSNAME keyword on the FRRECOV command. All flavors of data sets are supported. Multivolume, etc. We do a physical DSS copy (disk) or restore (tape) and then HSM recatalogs all of the pieces after the physical data recoveries complete. Recovery from tape is slow because DSS has to search the dump volume to find the extents. 2) You know the drill :) Open an RFE. With our crazy world, we have an extensive DFSMS roadmap around cyber resiliency enhancements. Stay cool :) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Flashcopy for Non-SMS volume backup?
Hey Ed, Chatted with some DFSMS folks about what you want to do. First, you are correct. The HSM FR support is only for SMS-managed volumes. It is tightly integrated with several SMS constructs for managing the volumes and policy. (In the future, feel free to send me a direct email to save yourself alot of reading :) Next, for your scenario, we that it will work if you do the following: 1) Identify your source to target pairings 2) We HIGHLY recommend that you use FC Incremental because it creates a persistent relationship between the source and the target. Because the HSM dump will be done in the standard dump path, it doesn't have the 'smarts' to do things like withdraw a NOCOPY relationship, etc. Also, more later on why it will be very useful. 3) Create the FC with DSS and dump conditioning to the target, and with FCINCREMENTAL 4) I believe that you are all set with adding the FC Target volumes to HSM, setting up the appropriate dump class, dump cycle, etc and having them picked up as a part of autodump / dump command. 5) What makes the dump copies fundamentally different from a standard HSM Dump or an FRBACKUP DUMP - a) DSS is smart enough to know that a FC dump conditioned target volume is being dumped, so the dump copy that it creates will be written with the volser of the source volume and will be created as if it came directly from the source b) HSM is smart enough to know that it is a dump conditioned volume, so the HSM CDS dump record and LIST output will record it as a dump of the target, with the FC target volser. Hence, the recommendation for a persistent relationship between your source and target volumes. Then you can manage the inconsistency between the HSM dump records and which volume was actually dumped. 6) As you indicate, you can use HSM Recover with TOVOLUME to recover directly back to the source. Since the dump volume is actually of the source, all should be well after the recovery ends. The volser will stay the same during the recovery process since the dump tape has the source volser. We believe that you shouldn't have to do a clip or anything after the Recovery. All should be well. There's a lot going on here. If you'd like to set up call to discuss, reach out and I can setup a call with all of the right SMEs to walk you through this. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM Backups using DFDSS - Enqueues
Hi Terri, You have the correct HSM setting - INUSE(...) HSM traps on the ADR412 message to know that DSS failed serialization and that it should be retried. Do you see a subsequent DSS Dump for the data sets that initially failed (a retry w/o serialization)? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
New online z/OS DFSMS Community
We're happy to announce the creation of the online z/OS DFSMS Community, which is part of the larger IBM Community! Our community webpage contains blogs to advertise new function (z/OS Releases and Continuous Delivery), advertise upcoming events (SHARE, IBM TechU, DFSMS Academies, etc), have Discussions (currently we're asking for input on topics for the next DFSMS Academy) and a library of helpful documentation. Join us today! https://community.ibm.com/community/user/ibmz-and-linuxone/groups/topic-home?CommunityKey=3c6c69c1-abde-4ee6-b603-19ef14942e8f -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Using HSM Transition to clear MOD-3 devices
Hi Chuck, We have clients who regularly use this function for what you plan on doing. I'll reach out to the individual who owns the ticket and make sure that it makes progress. Thanks. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: AWS down again.
DFSMShsm Migrate/Recall and DFSMSdss use the DS8000 to offload data directly to cloud object storage. S3 compliant object stores, including AWS, are supported. When the object store is unavailable for any reason, you'll get user friendly messages from HSM/DSS that indicate the specific error being returned by the object store provider. For HSM automatic migration, any data sets that failed to migrate will be reattempted the next day. The big concern with the unavailability is retrieving your data, which is not possible when the object store is unavailable. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm SETSYS BACKUP(DASD)
Hi Theo, HBACK is a data set backup command. Data set backup settings are determined by the SETSYS DSBACKUP parameters, of which the default for TAPE tasks is 2. Please update your SETSYS DSBACKUP parameters such that all backups will go to disk. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFDSS support for ZFS files query
For those following this thread who have a DFSMShsm license, HSM does support this. It can backup all files, files based on wildcards and exclude specific files as directed by an EXCLUDE. A UNIX shell command is also available, to perform the HSM backups/recoveries directly from UNIX. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: zSeries and using cloud for backups
The TS7700C provides a really nice option for what you are being asked to do. You can continue to take the backup copies that you do today, and have the TS7700 seamlessly migrate them out to the cloud for you. Best of both worlds! If you want to go directly to the cloud, the DFSMShsm/dss, DS8K transparent cloud tiering provides the value that 100% of the data movement is performed by the DS8K. None of the disk data passes through Z! DFSMShsm TCT migration has been available for some time. This gives you the capability to create a fast PiT backup on disk with DSS (with FlashCopy, all of the data movement is within disk controller), and then have HSM automatic migration move them to the cloud with tct. So, you get PiT backup copies in the cloud, with none of the data passing through Z. Also, DSS just announced this capability for Full Volume Dump and HSM has a statement of direction to provide the same for it's dump capabilities. DS8K TCT now supports compression when targeting a TS7700. Compression when targeting cloud object storage is in the short term roadmap. The DS8K natively encrypts the data, so that you have the piece of mind that the data was encrypted on your Z platform before it even lands on the ip connection to the cloud. Cloud Tape Connector is another nice IBM option that allows you to spin off a tape copy to cloud. Reach out to me if you would like any more information on any/all of the options and I can connect you with the folks to get you more information. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: PDS/e Encryption
DFSMSdss supports host-based encryption on it's dump command. The support was added before tape encryption became generally available. Still a possible solution for you since your tape environment doesn't support encryption. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Storage & tape question
Hi All, I want to give another perspective on the need for backup copies. The focus here is on physical loss of storage. With replication, and many clients having 2, 3 and even 4 sites, the probability of needing a backup copy to recover from a physical loss of data really has decreased. (Still there, none the less). BUT, the probability for logical data corruption has INCREASED. Accidental and malicious data corruption is instantly mirrored to all replication copies, making them useless. Working in HSM, I regularly see calls requesting assistance in recovering large amounts of data from backup copies. We're all human and we all make mistakes. Some of those mistakes result in data loss. Also, all products have programming defects and some of those defects result in data loss. This speaks nothing to the current environment where governments are mandating policies and procedures for protecting against malicious data destruction. Your only hope for recovery is a PiT backup prior to the data loss/corruption. Not all loss/corruption will be found immediately. So, your ability to recover is a factor of how long it takes you to determine that there was corruption/loss and how much your willing to invest in keeping backup copies for at least that long. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Storage & tape question
Hi KB, IBM-MAIN is great. There are shops that are all primary disk. Just make sure that you are aware of all of the considerations before going down that route. (The ones that I'm aware of are still HSM users to take advantage of ML1, class transitions, and backup). A few responses to your follow-on questions... Hi, Thank you for the detailed response Glenn, IBM-MAIN is truly amazing. > Migrate/Archive > The three purposes of HSM migration are to 1) compress the data so that the > footprint is smaller, 2) move it to a lower cost media so that the TCO is > lower and 3) move the data to an offline media that doesn't consume online > UCBs. When considering bringing all of your data back online, you need to > consider the impact of all three. 1) Assuming 3:1 compaction, you'll need 3x > the online storage. With zEDC, that will vary on both what you can get on the > primary storage and the ML2 storage. 3) For larger shops, the number on > online UCBs is a factor. It's not a factor for smaller shops. - 1) Compression - wouldn't it be enough to rely on z15 on-chip compression + the compression/dedupe done by the storage array itself? Sure, it may not be 3:1.. but worth evaluating? If the array itself is doing C+D, then "rehydrating" the data isn't a problem I believe? >> Glenn: z15 compression can be utilized for nonVSAM, not VSAM. >> Compression/Dedup are generally provided by offline storage systems or those >> that emulate them (like virtual tape). Dedupe is generally too slow for >> primary storage. I'm not aware of compression capabilities on primary >> storage. If you are currently deduping your backup copies, then they will >> take even more storage when they are moved to primary disk unless you can >> utilize a backup solution that does software based dedup and targets disk. 2) It's not just the storage cost though right.. (cost of a bunch of disk, S&S) vs (cost of tape emulation, physical carts, bandwidth, S&S, HSM's direct & indirect costs) >>Glenn: agreed. TCO has to be considered 3) Ok, the UCB thing can be problematic for big shops, agreed. There's only so much you can do with 3390-54 (are bigger volumes coming anytime soon?). >>Glenn. DFSMS currently supports EAV 1TB volumes. > Another thing to consider with an all disk environment is your 'relief > valve'. It's simple to migrate data to tape as a means of ensuring that you > always have enough space on your primary storage for growth. If you only have > primary storage, what is your exception handling case when you have > unexpected growth and no migration tiers to which to move your inactive data? > How quickly can you bring more primary storage online? Sorry, I know it sounds silly when I keep saying 'assume x/y/z is already catered to', but ... assuming primary storage provisioning is no longer a problem (apart from the UCBs mentioned above). > Another option is DS8000 transparent cloud tiering. This enables you to > migrate inactive data to cloud object storage, with minimal cpu since the > DS8K is doing the data movement. If not a primary means of migrating data, it > is a very good option for a 'relief valve'. Hmm... the two whole approaches (all-primary vs standard procedure) need to costed out and compared to be impartial to either case. > Backup > Regardless of the replication technique that you are using > (synchronous/asynchronous), you need point-in-time copies of your data for > logical corruption protection. If a data set is accidentally or maliciously > deleted, replication quickly deletes it from all copies. Also, if data > becomes logically corrupted, it is instantly corrupted in all copies. So, you > have to have a point-in-time backup technique for all of your data. You need > as many copies as you want recovery points. One copy doesn't give you much > security. Keeping n copies on disk can get pricey and consume alot of > storage. Also, you need to replicate the n PiT copies to all of your sites so > that you can do a logical recovery after a physical fail over. This makes the > cost add up even more quickly. TCT is another good option for this. You can > keep 1 or 2 copies on disk and then have HSM migrate/expire the older backup > copies to cloud object storage which is then available at all of your > recovery sites. If we consider that the storage array has *proper* support for multi-site, snapshots/PiTs, etc. ... again not problematic. Fully understand I may be dreaming about such a primary storage, it's good to know the technical constraints against it. - KB -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Storage & tape question
A few thoughts: Migrate/Archive The three purposes of HSM migration are to 1) compress the data so that the footprint is smaller, 2) move it to a lower cost media so that the TCO is lower and 3) move the data to an offline media that doesn't consume online UCBs. When considering bringing all of your data back online, you need to consider the impact of all three. 1) Assuming 3:1 compaction, you'll need 3x the online storage. With zEDC, that will vary on both what you can get on the primary storage and the ML2 storage. 3) For larger shops, the number on online UCBs is a factor. It's not a factor for smaller shops. Some clients have selected to go to an all HSM ML1 environment to still get the advantage of zEDC compression on inactive data. (You may be utilizing zEDC for primary storage, but that is only available for nonVSAM data). These clients utilize the lowest cost disk and utilize the value of zEDC compression to minimize the footprint. Another thing to consider with an all disk environment is your 'relief valve'. It's simple to migrate data to tape as a means of ensuring that you always have enough space on your primary storage for growth. If you only have primary storage, what is your exception handling case when you have unexpected growth and no migration tiers to which to move your inactive data? How quickly can you bring more primary storage online? Another option is DS8000 transparent cloud tiering. This enables you to migrate inactive data to cloud object storage, with minimal cpu since the DS8K is doing the data movement. If not a primary means of migrating data, it is a very good option for a 'relief valve'. Backup Regardless of the replication technique that you are using (synchronous/asynchronous), you need point-in-time copies of your data for logical corruption protection. If a data set is accidentally or maliciously deleted, replication quickly deletes it from all copies. Also, if data becomes logically corrupted, it is instantly corrupted in all copies. So, you have to have a point-in-time backup technique for all of your data. You need as many copies as you want recovery points. One copy doesn't give you much security. Keeping n copies on disk can get pricey and consume alot of storage. Also, you need to replicate the n PiT copies to all of your sites so that you can do a logical recovery after a physical fail over. This makes the cost add up even more quickly. TCT is another good option for this. You can keep 1 or 2 copies on disk and then have HSM migrate/expire the older backup copies to cloud object storage which is then available at all of your recovery sites. Glenn Wilcock DFSMS Chief Product Owner -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm questions
First, I'm impressed that your subject line has 'DFSMShsm' as opposed to DFHSM or any of the other names that it's gone by over the years. Background on the design for selecting the data set backup target tier, first created when physical tape was still dominant - for small data sets, you don't want the overhead of a tape mount for a one-off backup, and for large data sets you don't want them going to disk and eating up all of your space. So, we created a SETSYS called DSBACKUP where you could fine tune just which data set backups go to disk and which go to tape and the tasking level for each. With virtual tape, this is no longer quite an issue, but the parameters still persist. If you always want you individual backup copies to go to disk, you can set the parameters to do so. In the DFSMShsm Storage Administration manual SETSYS chapter, there is a section names 'SETSYS fast-path for DFSMShsm Functions'. In this section you will find a table with the parameters that affect tape processing. But, Chapter 10 of the DFSMShsm Implementation & Customization Guide is an exhaustive guide for establishing the DFSMShsm tape environment. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMS Move Dataset
If you have DFSMShsm, HSM has a simple MIGRATE DSNAME(dsname) MOVE command that will invoke SMS ACS for volume selection and move the data set (any type) to a new volume. Under the covers, HSM invokes DFSMSdss COPY w/ DELETE, w/o you having to create all of the JCL based on data set type. It can also be issued at the VOLUME and STORAGEGROUP level to drain a volume or storage group. For example, it's a simple single command that can be used to migrate from smaller volumes to newly defined larger volumes. It will even close open Db2 objects, move them and then automatically reopen. https://www.ibm.com/support/knowledgecenter/SSLTBW_2.4.0/com.ibm.zos.v2r4.arcf000/s4388.htm -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm high CPU consumption [EXTERNAL]
Hi Lizette, It allows different WLM settings. For example, a MAIN can do RECALLS and have a higher WLM goal than an AUX host that is assigned to run a migration window with a lower goal. I've never actually worked with a client to do this... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm high CPU consumption [EXTERNAL]
It is Newsletter 2018 No. 4 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm high CPU consumption [EXTERNAL]
DFSMShsm has made many advances over the years to become more efficient. I teamed up with Frank Kyne of Watson & Walker to write an article on this topic in their newsletter. I'm also more than happy to give a WebEx on this topic with this client, and any other who is interested. It's a popular presentation from IBM TechU and SHARE. One of the topics is to reduce the MAX tasking level to reduce peak MIPS during HSM windows. Just send me an email. Glenn Wilcock, DFSMS Architecture wilc...@us.ibm.com -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM/DCOLLECT QUESTION [EXTERNAL]
HSM's usage of RLS Access for the control data sets is driven by the PROCLIB value for CDSSHR. CDSSHR=RLS indicates that HSM should access the CDSes in RLS mode. If CDSSHR is not specified, or has any other value, then you are not using RLS for HSM. DCOLLECT behaves this way because it is not running under the HSM address space, so it is how we determine the proper way to open the control data sets concurrently with HSM. BTW - Using RLS to access the HSM control data sets is my second rated best practice for HSM. If you run concurrent HSM activity across multiple LPARs, it can significantly improve elapsed times. My highest rated best practice for HSM is to use zEDC. zEDC does wonders for HSM elapsed times and storage reduction. If migrating to ML1, it also reduces cpu consumption. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
The Next Evolution of Hierarchical Space Management
The next evolution of Hierarchical Storage Management is here with the announcement of TS7700 DS8000 Object Store! This evolutionary capability enables DFSMShsm to migrate data directly between IBM DS8000 Disk and TS7700 Virtual Tape without the data passing through the Z host! Check out more at the IBM z Systems Development Blog: https://www.ibm.com/developerworks/community/blogs/e0c474f8-3aad-4f01-8bca-f2c12b576ac9/entry/The_Next_Evolution_of_Hierarchical_Storage_Management?lang=en -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
DFSMS Continuous Delivery
Two recent z/OS DFSMS Continuous Delivery projects: - OA57454 Added DELETE (Archive) capability to HSM File Level Backup Support (base support in OA52703) - OA57173 Added initial support for FlashCopy to Global Mirror Primaries (https://www.ibm.com/developerworks/community/blogs/e0c474f8-3aad-4f01-8bca-f2c12b576ac9/entry/FlashCopy_onto_Global_Mirror_Primaries?lang=en) -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Silly DFSMShsm questions
If it happens again, open a pmr. ISPF uses device type to determine if MIGRAT'x' corresponds to ML'1', ML'2' or ML'C'. Since you are seeing a 'C', but have never targeted 'C'loud object storage, there could be a defect in which the device type is not being set appropriately. Thx. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Stay current with what's new in DFSMS
Certainly. We are just trying to utilize new avenues for disseminating information. One size never fits all. In a recent LinkedIn post, I documented all of the DFSMShsm RFEs that are included in V2R4: RFE 121915: (SHARE MVSS) Quiesce window for interval and on-demand migration Request: HSM automatic migration via interval or on-demand migration sometimes runs during peak CPU utilization times. We are looking for the ability to provide windows in which we can turn off both types of automatic migration. It is already disabled during primary space management (PSM) processing; we look for additional times to disable it. We want to keep command migration available at all times. Disabling automatic migration should not prevent command migration. Response: New SETSYS EventDrivenMigrationQuiese and HOLD/RELEASE EventDrivenMigrationQuiese commands. RFE 114044: HSM should include which parameters are valid/invalid during startup Request: When DFSMS starts up, it produces messages on parms in ARCCMDxx either being valid or invalid. It would be good to know which parm was verified (or not). So it could be ARC0100I SETSYS COMMAND COMPLETED for - X or ARC0104I INVALID INITIALIZATION COMMAND - Response: Provided RFE 130361: Empty data set during backup flag Request: In the Data Areas manual there is MCDFMPTY field that when set to 1, the data set was empty at the time of migration. Would like the same for a backup. Response: MCCFMPTY added. The following messages are now issued to SYSLOG to enable automation: RFE 116187: ARC0019I CELLPOOL cellpool ENCOUNTERED A SHORTAGE RFE 126300: ARC0263I DUMP VOLUME volser DELETED now issued to SYSLOG RFE 132117: ARC1814I FAST REPLICATION BACKUP HAS COMPLETED SUCCESSFULLY AND DUMP IS NOW STARTING RFE 119146: Improve FREEVOL messaging Request: During FREEVOL of an ML1 volume, if an ARC0560E condition exists where migration to ML2 is limited due to lack of ML2 tape/DASD, datasets that should be migrated to ML2 stay on the ML1 volume. The corresponding ARC0734I message indicates that the dataset moved "TOVOL **", but the return code is zero and the reason code is zero. They should be RC5 REAS8. Response: Provided RFE 115386: ARC1001I messages routed to activity log Request: BACKDS fails, we would like to see the failing ARC1001I message issued to the backup activity log. Response: Provided -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Stay current with what's new in DFSMS
The DFSMS team recently created a LinkedIn Group for z/OS DFSMS: https://www.linkedin.com/groups/12238880/ Join the group to stay current on the latest enhancements, tips, news, etc. For example, while z/OS release RFAs describe all of the major enhancements in a release, they don't include all of the minor RAS items and RFEs. In my latest post to the group, I list all of the DFSMShsm RFEs that are included in V2R4... with more to follow later in the year. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Thoughts on IBM z/OS Version 2 Release 3 enhancements
Documentation link is now live: http://publibz.boulder.ibm.com/zoslib/pdf/OA52703.pdf -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Thoughts on IBM z/OS Version 2 Release 3 enhancements
We just finished up an early test with a handful of clients and are in the process of closing all of the V2R3 APARs. (Due to preReq's, they all can't close at the same time). The plan is for all of the ptf's to be available in May. The support will be in the base of V2R4. Thanks, Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM management
Have been wanting to add Query ODS for a while, along with making QUERY it's own task. RFE 65675 requests this. The more votes, the better the odds... Regarding the main thread, the REPORT command is limited, but the Report Generator support enables customized reports of the HSM SMF and CDS records. ISVs do offer very nice reporting/management capabilities. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM and empty files
Hi, Sorry to hear that you feel this way. Our desire is to balance providing leading edge solutions while meeting our clients' needs for base functionality. Recently, the z/OS team switched to using RFE so that we can better manage client requests. It enables clients to vote on enhancements that are most meaningful to their environments. If there is a specific improvement that you'd like to see, please open an RFE and we can rank it with the other RFEs that we receive, based on overall client interest. In the last couple of years, the HSM team has closed 18 unique RFEs (a total of 23). The aforementioned APAR to process data sets with undefined data set orgs being one of them. Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM and empty files
Hi, I looked and didn't see any Patches for skipping empty data sets. HSM treats these as normal since both migration and expiration can depend on the existence of a valid backup copy. For this reason, HSM recently added support for data sets with an undefined data set org (OA52304), since some clients reported that a large number of these data sets were consuming a large amount of primary storage and HSM was not able to process them. Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM Migration Question
HSM uses the Base Data Component to determine data set migration eligibility. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMSHSM, RMM - Setup
Hi Jake, Redbooks are always a great place to start to learn the basics. Both DFSMShsm and rmm have primers. Refer to the PROCLIB and PARMLIB members of each to understand how they are each currently configured. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM options for dataset back-up
DFSMShsm doesn't have any documented way to do this, and I haven't found any undocumented patches either. DFSMShsm also provides a NORESET option to not reset the DSCI when doing full-volume dumps. I suspect that this is common across all backup products for full-volume processing, so that data set backups can continue as normal regardless of the dump cycle. Not sure of your specific scenario, but look at SMS GUARANTEED BACKUP FREQUENCY. It drives periodic backups even if the DSCI is not set. It has some caveat's, but it's worth looking at if that is what you are trying to solve. While we always welcome RFEs, just want to say up front that unless there are many additional votes from other clients also requesting this type of enhancement, it's unlikely that a unique request like this will ever make it high enough in our backlog that it will get done. Glenn Wilcock DFSMShsm architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Auto backup of zfs
Hi, we have not yet publicly announced a date. If you have a nonDisclosure Agreement, we can discuss offline. Thx, Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Auto backup of zfs
Please refer to this Statement of Direction regarding the intent to enhance DFSMShsm and DFSMSdss to backup and recover individual files within a zFS. https://www-01.ibm.com/common/ssi/cgi-bin/ssialias?infotype=AN&subtype=CA&htmlfid=877/ENUSZP18-0290&appname=lenovospain&language=es Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: How to exclude files from HSM Backup or Migration
A recent enhancement enables HSM size thresholds to be specified within a management class. For your specific case, you can specify within the management class that data sets > xx in size should not be migrated. (For those who are interested, yes, you can also specify that data sets <= xx in size should not be migrated, or data sets > x in size should always go to ML2, etc. Gives you some flexibility to do things based on size in the management class instead of having to write up an MD exit). The APAR is OA52901. Unfortunately, it was marked PE because of a MULTISYS issue. The PE fix should be available in the near future. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
What do Space Invaders and DFSMShsm have in common?
What do Space Invaders and DFSMShsm have in common? They share an original product release year of 1978! Back when a gallon of gas was 63 cents, the DOW closed at 805, Illinois Bell introduced the first cellular mobile phone and Saturday Night Fever was all the rage. As with everything else, DFSMShsm has transformed immensely over the last 40 years. At the upcoming SHARE in Sacramento, we'll be celebrating DFSMShsm's birthday at the session "DFSMShsm at 40: A Lean, Mean, Data Management Machine"! At this celebration, I'd like to highlight some of the HSM record holders that we have out there. Please respond to me with the values for your shop. Please reply to me directly as opposed to posting back on this board. At SHARE, I'll announce the winners. I will leave off company names / responder names unless I get an explicit permission to include them (excluding the first question). To get the information, use DCOLLECT for the HSM records and apply a Sort. Better yet, the add-on HSM products should also return this information. - When did you first start working on DFSMShsm (even if you don't work on HSM anymore)? (I plan to have a prize for the person present at SHARE who has worked on HSM the longest) - What is the creation date of your oldest migrated data set? (05 Nov 1987 is the earliest date that I have so far) - What is the creation date of your oldest backed up data set? - How many tapes does your DFSMShsm manage? (Total tapes defined to HSM, regardless of logical or physical) - What's the total number of migrated data sets that you have? - What's the total number of backed up data sets that you have? Thanks! I look forward to seeing you at SHARE! Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFSMShsm Copy Pools
Hi, There are many large clients who use DFSMShsm copy pools to backup their DB2 systems. Some also use it to create offline, full-volume copies of other volumes. Only the FRBACKUP/FRRECOV/FRDELETE commands operate against copy pools. FRRECOV does have a data set recovery option. I recommend referring to the following Redbook: DFSMShsm Fast Replication Technical Guide. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM QUESTION
Correct. 'MCDS' is the default. You just need to specify one of the backup keywords. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM Backup to Disk
VTS does wonders for HSM. Others have made good points about having all of the backup copies on disk. One other thing that I would like to point out would be the increase in CPU consumption and elapsed time. With tape, the compression is done within the tape controller. If you are backing up to disk, HSM must do the compression. That will significantly increase the CPU and will probably elongate the elapsed time. If you do go with disk, I would highly recommend using zEDC for the compression. That will actually decrease the CPU and improve the elapsed time because the compression ratio is so good and the data is compressed by DSS before being passed to HSM. Cloud Tape Connector is not an option for HSM data. CTC works on SMS tape data sets, of which HSM and OAM data are not. As noted earlier in this same thread, HSM stacks the data on tape as a single tape file and internally manages the location of each logical data set on tape by storing the File Block ID. CTC doesn't work with this type of setup. Regarding Cloud Storage, HSM and DS8K recently announced support for Transparent Cloud Tiering where the DS8K directly writes migration data to cloud storage. The storage can either be on prem or off prem. The value being that the data no longer flows through the z server, significantly reducing the CPU requirements and eliminating RECYCLE processing. But, backup support is not yet available. For those concerned about the data going to 'the cloud', we expect most z data to go to an on prem cloud, which is within the walls of your existing environments. When going to an off prem cloud, z/OS will be providing data set level encryption, so that the data is encrypted on z before being sent out. By policy, data will be able to be sent to ML2, on prem cloud or an off prem (public or private) cloud. Additionally, IBM announced limited support for this same cloud support to target an IBM VTS. This enables direct data movement between an IBM DS8K and IBM VTS for HSM migration and recall processing, also enabling significant CPU reduction, elimination of 16K block writes and elimination of RECYCLE processing. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Ransomware on Mainframe application ?
Pointers to some information from IBM on this topic... https://securityintelligence.com/ is an IBM site that provides security related insight & analysis and provides valuable information regarding Ransomware, including WannaCry. https://www-03.ibm.com/systems/z/solutions/security_integrity.html provides an overview of z Systems Security and is intended to help you stay current with security and system integrity fixes by providing current patch data and associated Common Vulnerability Scoring System (CVSS) ratings for new APARs. Security Notices are also provided to address highly publicized security concerns. This link also includes the IBM z/OS System Integrity Statement, a portion of which states "IBM’s long-term commitment to System Integrity is unique in the industry, and forms the basis of z/OS’ industry leadership in system security. z/OS is designed to help you protect your system, data, transactions, and applications from accidental or malicious modification. This is one of the many reasons IBM z Systems remains the industry’s premier data server for mission-critical workloads." In addition to preventing Ransomware, enterprises need to protect data from being stolen. IBM issued a Statement of Direction in the Announcement letter IBM United States Software Announcement 216-392, dated October 4, 2016, communicating "IBM plans to deliver application transparent, policy-controlled dataset encryption in IBM z/OS®. IBM DB2® for z/OS and IBM Information Management System (IMS™) intend to exploit z/OS dataset encryption." When recovering from an accidental or malicious data destruction event, z/OS DB2 provides the RESTORE SYSTEM utility that recovers a DB2 instance to a specific point in time before the data was destroyed. Additionally, IBM is gathering client feedback regarding extending this functionality to an entire enterprise. Glenn Wilcock DFSMS Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM followup question
HSM uses the current values of the assigned management class to manage a data set. So, any changes made to the values of a management class will be used the next time that HSM selects for processing any data sets assigned to that management class. A data set may be assigned to a different management class. In fact, in z/OS V2R1 and higher, the HSM class transition support can be used to change the management class, storage class and/or storage group based on policy (note that data class is excluded). In z/OS V2R2, the HSM MIGRATE TRANSITION command can also be used. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: CREATED date for migrated data sets?
Specifically, MCDDLC (Creation date stored in MCD) is copied to UMCREDT in the 'M' record for DCOLLECT. Further, the DFSMS Report Generator function can be used to process DCOLLECT M records to pull just the records that you want. Refer to Chapter 14 of the DFSMShsm Storage Administration. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: EAV bug or feature?
Besides the Redbooks, a good place to start is the 'z/OS DFSMS Using New Functions' manual. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Example of ACS Environment of SPMGCLTR
Hi Theo, Yes, SPMGCLTR was introduced in the ACS routines in 2.1. It is used by HSM for transitions for structured data, opposed to CTRANS which is used by OAM for unstructured data. I was incorrect in my earlier post. After communicating with the developers, they indicated that management class, storage class and storage group should be passed into the SPMGCLTR routine as read only variables, but those are the only ones. We need to add additional variables like UNIT, etc. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Example of ACS Environment of SPMGCLTR
Hi Theo, Change: IF &ACSENVIR = 'SPGMCLTR' THEN SELECT (&MGMTCLAS) WHEN ('USRTIER1') SET &MGMTCLAS = 'USRTIER2' OTHERWISE SET &MGMTCLAS = &MGMTCLAS END to: IF &ACSENVIR = 'SPGMCLTR' THEN SET &MGMTCLAS = 'USRTIER2' END This obviously overly simplifies it, unless you only plan on making the change to one management class. Unfortunately, with the current limitations, you may not have access to the RO variables that you need to add the correct logic. I also forgot to point out that we shipped a Development APAR OA47700 that enables you to override ACS logic and just specify MC, SC and / or storage group on the various Migrate Transition commands. For example, you could transition a data set to just a new storage group with the command HMIGRATE DSNAME(... ) TRANSITION(STORAGEGROUP(Nearline)). If you'd like to set up a conference call to discuss your options in more detail, you can email me at wilc...@us.ibm.com. Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Example of ACS Environment of SPMGCLTR
Hi, I believe that the problem is that my original presentations on this topic showed using the existing classes/group to determine the new class/group. Unfortunately, that was function that did not make shipped support. (I have updated the examples in my most recent presentations). For clients who need that support, I have been requesting that they open an RFE. We definitely have on our roadmap the addition of that functionality. Since the existing management class is not being passed in, your logic that sets the new management class based on the current management class is skipped, meaning that the management class is not changed. Same for the SC and SG. Since none of the classes/groups change, that is why you are getting the message that you are. A data set is only eligible to transition if at least there is one change. I believe that if you remove the check for the existing class and just set the class/group to a specific value, that it will get around this problem. Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM QUESTION
You are correct. 'PREPARE' only creates an ABR record. No verification or data movement occurs. Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM/SMS QUESTION - SPACE MANAGEMENT
Hi, a couple of other things to look at: (1) Verify that the volume is being selected for space management. As indicated, HSM doesn't process data sets on a volume unless the volume exceeds its high threshold. Once a volume is selected, HSM will process all eligible data sets on that volume until the low threshold is reached. (2) Enable PATCH .MGCB.+26 X’FF’, which indicates that HSM should issue additional ARC0734I messages to indicate why data sets weren't selected for processing. (This patch is documented in the DFSMShsm Diagnosis manual). Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: RLS implementation for CDS's in DFHSM
Our top recommendation for improving DFSMShsm throughput is to use RLS for the HSM CDSes. In nonRLS mode, each HSM locks access to the CDSes while it performs its queued up I/O requests. While that host has the CDSes locked, all other HSM hosts are queueing up their I/O. When a host gets exclusive access, it has to first flush all of its buffers because they are no longer valid and then do a VSAM VERIFY. All of the buffers are then repopulated for the I/Os that have been queued up. All of that overhead goes away with RLS. Also, Read I/Os are faster, so read intensive workloads like ExpireBV, Secondary Space Management and Audit perform faster, even when there isn't alot of multiple host concurrent workload. The higher the amount of concurrent HSM host activity that you have, the greater the improvement that you'll see with RLS. In z/OS V2R1, DFSMShsm enhanced its error recovery for SMSVSAM server errors. Prior to V2R1, when there was an SMSVSAM server error, HSM lost access to the CDSes, so HSM would just take a fatal abend. (Yes, very ugly, we know). Starting in V2R1, DFSMShsm will hold I/O when it identifies that there is an SMSVSAM error. HSM waits for the SMSVSAM server to restart. After the server restart, HSM will automatically retry all of the I/Os that were held. The only I/Os that have to be failed are Update I/Os after a Read for Update, when the Read for Update was before the server error. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM Recovery question
Hi, Did you happen to see any errors with reading the VTOC copy data set? From the Storage Admin: "DFSMShsm uses the latest VTOC copy data set to recover the volume unless an error occurs in allocating or opening the latest VTOC copy data set. If the error occurs, DFSMShsm attempts to recover the volume, using the next latest copy of the VTOC copy data set. If a similar error occurs again, the volume recovery fails." You indicate that you only have one VTOC copy. Does the date correspond to the Week 2 Dump date? I wasn't able to find any other logic that would cause HSM to recover data from the Week 1 VTOC copy. If the above don't apply, then a pmr could be opened to analyze the situation. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM Recalls and MASKING
Hi Lizette, We want to have the documentation as helpful as possible. It would certainly be reasonable for you to open a Request for Publication Change to add information in the HSM SA to indicate that the DFSMShsm Managing Your Own Data has additional information available regarding using wild cards when a user is recalling data. Thanks, Glenn Wilcock -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM - Dump w/tolerate enqueue failure ?
Hi, If you have FlashCopy, you can create Point in Time backup copies and then dump them to tape. This is supported via the DB2 BACKUP SYSTEM utility and DFSMShsm FRBACKUP. You can also use DSS to create Image Copies. I think that it would be best if we have a call with you, DB2, DSS and HSM (myself) to discuss the various options available to backup DB2 data. I'll set up a call. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM - Dump w/tolerate enqueue failure ?
Hi, DFSMShsm invokes DSS DUMP FULL to create dump copies. HSM specifies the TOLERATE(IOERROR) keyword. This varies from the ENQF option. ENQFAILURE is mutually exclusive with FULL. TOL(IOE) indicates that I/O errors on the volume being dumped should not fail the dump (unless there are 100 or more). On restore, the tracks that couldn't be read are cleared. (Refer to the DSS manual for DUMP FULL and TOLERATE(IOERROR)). Full volume dump processes at the Track/Cylinder level, not the data set level. Individual data set enqueues are not obtained. The VTOC is locked to prevent any creates, extends, renames and deletes during the dump. Active data would be a Fuzzy backup if it is not being quiesced during the dump, but it would still be dumped. What are you seeing when you indicate that no data is backed up? Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM full vol dump question
Hi, You can also try LIST PRIMARYVOLUME ALLDUMPS BCDS. (You may have to specify a specific volser). Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Thoughts on DFSMShsm ABARS in today's environment
It uses IBM FlashCopy, which both Hitachi and EMC support. (We have clients with a mixed vendor storage environment who have implemented this solution). Glenn -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Thoughts on DFSMShsm ABARS in today's environment
Hi Again, As you can imagine, over the years our L2 folks have been called out to help recover data after many an 'oops'. In many cases, there isn't a backup for all of the data, which has resulted in data loss. Hence, most will concur that any backup is better than no backup. For the case of backing up Online Systems, investigate what the applications themselves offer. In the case of DB2, DB2 and DFSMS have worked together to create a Continuous Data Protection solution (CDP) through the DB2 BACKUP SYSTEM / RESTORE SYSTEM utilities. BACKUP SYSTEM creates a fuzzy backup that can be made consistent during recovery through Log Applies. In order to create a consistent recovery, creates/renames/deletes/extended must be quiesced during the backup, but data updates are allowed to continue. (Essentially VTOC updates are held to ensure that the VTOCs are data consistent). DB2 invokes DFSMShsm to create FlashCopy PiT backups (FRBACKUP). During Recovery, the client selects the recovery point, which can be 'current', or a prior point in time. DFSMShsm recovers the PiT copy using FlashCopy (if still on disk) or from Tape (if the backup copy has rolled off to tape). DB2 then performs a log apply up to the recovery point, creating a data consistent recovery. Some clients also use this process to create Clones. When this CDP solution is combined with Metro Mirror, there is a very high level of protection from both physical and logical loss. This solution can also be used for SAP systems, when DB2 is the database. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Thoughts on DFSMShsm ABARS in today's environment
Hi Lizette, There have been some great appends to this question. I also wanted to add my two cents... When determining your backup strategy, you need to determine the desired RTO (Recover Time Objective), RPO (Recovery Point Objective) and BWO (Backup Window Object) for both physical loss (storage failure) and logical loss (data corruption) at an acceptable price point. This can be daunting. You'll find that this may require that multiple backup strategies be used for the various types of data in your environment. To protect against physical data loss (some type of storage failure, whether it is a result of a failure with the storage itself or an external cause such as power outage, natural disaster, etc), there are many types of mirroring solutions. These solutions create two or more copies of the data at different locations such that if the primary storage fails, processing can switch to the alternate location. To protect against logical data loss (some type of data corruption, whether it is caused by a software defect or human error), there needs to be point-in-time copies. Mirroring solutions do not provide any protection here because the logical corruption is just instantly mirrored to all locations of the data. There are many different types of PiT copy techniques. One of the important factors is to ensure that the data is quiesced so that it is a data consistent backup. Or, that the recovery technique is able to create a consistent recovery from a fuzzy, non consistent, backup. The Use Case for ABARS is when you need to create a data consistent backup of a set of data that can be used to recover that entire set of data to the same point in time. As was mentioned in another append, an example is backing up an entire Payroll application that can be recovered in it's entirety with a single recovery command. After 9/11, I visited a client who was setting up ABARS just for Payroll because they wanted to guarantee that they could pay their employees after a disaster... If you want you employees to show up after a disaster... you better be sure that you can pay them. (In this case, ABARS was being used to protect against physical data loss). ABARS also allows you to recover one or more individual files from the aggregate backup. When mirroring alternatives are too expensive, ABARS can be used as a lower cost DR solution that clients use as follows: For backups, full volume dumps are created for the system packs and ABARS is used to backup all of the various applications and their related data. DR tests are performed by renting space at an offsite facility. The first step is to recover all of the system packs and start the system. They then recover the remainder of their data by prioritizing the ABARS recoveries. They start with their most critical applications and bring them up as the recoveries complete. So, when determining your backup strategy, first determine your recovery needs and then determine which backup technique will best fit those needs. Glenn Wilcock DFSMShsm Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: DFHSM BCDS Z/OS 1.13
Hi, I agree that it just sounds like your CDSes have just grown and need more space. Please note that the CA Reclaim function became available in V1R12. When CA Reclaim is enabled for the CDSes, it minimizes the need to reorganize the CDSes, especially after running EXPIREBV. When a VSAM Control Area becomes empty, CA Reclaim will automatically free up that space, which is a primary need to run a reorg. Glenn Wilcock HSM Architect -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: HSM ML2 on disk
Hi, The HSM ML2 Disk support is very basic with no plans to enhance it. For clients who don't want physical tape anymore (for whatever reason), they generally select either a disk-only VTS ML2 or implement all ML1. If you go all ML1, I would recommend implementing ML1 Overflow volumes to help ensure that you have space for larger data sets. For ML1 Overflow volumes, the selection algorithm selects the volume with the least amount of freespace that will contain the data set, in order to maximize the most amount of freespace on the volumes. (Standard ML1 selection chooses the volume with the most freespace which results in volumes getting filled evenly but not leaving any volume with a significant amount of freespace for larger data sets). Evey client that I have talked to that has implemented a large disk-cache VTS is very pleased with the HSM improvements. For large customers and those who need to minimize storage costs, the VTS also has backend physical tape. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN