Re: Mod-9 vs. Mod-27 vs. mixed
Thank you all for your insightful user experiences in response to my question. If I may summarize the points that apply to my particular situation: * No real issues with respect to backups, particularly for consolidations, e.g. of volume pools -- in fact we may even see a slight improvement in backup turnaround. * No special considerations for mirroring, the same challenges would seem to apply in general, in all three cases. * The standard disclaimer still applies, however, with respect to critical small datasets, such as the JES2 checkpoint and combinations of couple datasets, that requires a compromise between simplifying the configuration and wasting a LOT of space on a MOD-27 or maintaining a small subset of MOD-9s and wasting much less space in a slightly more complicated configuration I should also have mentioned: No HyperPAV, possibly PPRC or maybe XRC around the corner, and all backups under the umbrella of DFSMShsm. Again, thanks -- IBM-Main rules! Tom Sims zOS guy since the daze of XA... -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
Don, I'm can't speak for the EMC and IBM iterations, but with HDS HDP setting up mirrored configurations is just way to easy. I'm guessing it would be similarly easy on the other vendors wide stripe configurations. It's not exactly what you are asking for, but in the lab we are setting up and tearing down remote copy configurations all the time. One day it's 3390-9 and the next day it is 3390-A fill size. The pools themselves remain the same, it's just a case of defining the new volumes in the pool, and growing/shrinking the pool as necessary. When finished you just delete the volumes and create new ones of whatever size you want. It's not the nirvana of a self-replicating configuration, but it's is far easier than the old days of reformatting arrays, or deleting and creating custom volumes in the middle of a parity group or looking for small leftovers at the end of a parity group. Ron -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Don Williams Sent: Friday, February 22, 2013 2:36 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: [IBM-MAIN] Mod-9 vs. Mod-27 vs. mixed I think DASD vendors need to create a simple, standard method to mirror a volume's configuration and data, rather than mirroring just the volume's data. Don -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Skip Robinson Sent: Friday, February 22, 2013 3:40 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Mod-9 vs. Mod-27 vs. mixed In the interest of frugality, I asked my storage guys some time ago to allocate some tiny volumes for JES checkpoint and couple data sets. After a while, they complained that it was more trouble than it was worth because we mirror most volumes to the DR site. For every tiny source volume, they needed a corresponding tiny mirror volume. Periodic DASD refresh (aka upgrade) projects only added to the complexity. Pennies vs. pounds. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Staller, Allan allan.stal...@kbmg.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 10:04 AM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM- m...@listserv.ua.edu Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed)
Ed, While it's not perfect, what if you did the FlashCopy to a completely controller? If you have HDS DASD, you could virtualize some brand-x midrange disk and FlashCopy or Shadowimage the zFS files/volumes to the midrange storage. Using FlashCopy Incremental means that you only copy what you change. If offsite backup is required then you can replicate the virtualized volume, or you can still backup the midrange volume to tape. It's going to take the same amount of time, but it would be out of the critical path. There's plenty of ways to skin a cat. This is just one way that I thought a HDS site could approach the problem of long backups of zFS with small change rates. Ron -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Ed Jaffe Sent: Saturday, February 23, 2013 9:45 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: [IBM-MAIN] Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed) On 2/23/2013 9:33 AM, Paul Gilmartin wrote: Most of my department's large UNIX filesystems are NFS mounted from ZFS (not zFS) on Solaris servers. Our daily backups are ZFS snapshots, almost negligible latency, followed by background dumps to tape. Right. FLASHCOPY of a large volume on our DS8100 can be done in almost an instant. But, copying the volume's data to tape takes hours -- as mentioned previously. My question is about whether a DFS/SMB ZFS should be backed up at all given its size and the existence of a daily TSM backup of its contents. -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed)
On 24/02/2013 4:45, Ed Jaffe wrote: My question is about whether a DFS/SMB ZFS should be backed up at all given its size and the existence of a daily TSM backup of its contents. In theory, a TSM backup of the contents should be adequate. In practice, there are a few things that I would want to be sure TSM would actually backup and restore before I considered it a full filesystem backup. I would want to be sure TSM handled correctly: - symbolic and hard links (i.e. hard links restored as entries for the same file) - special files - ACLs, file tags, audit bits, etc. In other words, does TSM back up the data, or the filesystem (data + metadata). Regards Andrew Rowley -- and...@blackhillsoftware.com +61 413 302 386 -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
My shop has IBM DS8000 DASD. Defining mirror configurations is easy, I just think it should be automatic. For example: 1. I cannot say define target volume like the source volume. 2. I cannot change the size of the pair of source and target volumes as a single operation. I have to break the relation, separately change the source and target, then re-establish the mirror relationship. Yes, I could write scripts to do that, but I think that these kinds of tasks are so basic and fundamental that they should be part of the base product. Don -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Ron Hawkins Sent: Sunday, February 24, 2013 5:07 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Mod-9 vs. Mod-27 vs. mixed Don, I'm can't speak for the EMC and IBM iterations, but with HDS HDP setting up mirrored configurations is just way to easy. I'm guessing it would be similarly easy on the other vendors wide stripe configurations. It's not exactly what you are asking for, but in the lab we are setting up and tearing down remote copy configurations all the time. One day it's 3390-9 and the next day it is 3390-A fill size. The pools themselves remain the same, it's just a case of defining the new volumes in the pool, and growing/shrinking the pool as necessary. When finished you just delete the volumes and create new ones of whatever size you want. It's not the nirvana of a self-replicating configuration, but it's is far easier than the old days of reformatting arrays, or deleting and creating custom volumes in the middle of a parity group or looking for small leftovers at the end of a parity group. Ron -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Don Williams Sent: Friday, February 22, 2013 2:36 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: [IBM-MAIN] Mod-9 vs. Mod-27 vs. mixed I think DASD vendors need to create a simple, standard method to mirror a volume's configuration and data, rather than mirroring just the volume's data. Don -Original Message- From: IBM Mainframe Discussion List [mailto:IBM- m...@listserv.ua.edu] On Behalf Of Skip Robinson Sent: Friday, February 22, 2013 3:40 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Mod-9 vs. Mod-27 vs. mixed In the interest of frugality, I asked my storage guys some time ago to allocate some tiny volumes for JES checkpoint and couple data sets. After a while, they complained that it was more trouble than it was worth because we mirror most volumes to the DR site. For every tiny source volume, they needed a corresponding tiny mirror volume. Periodic DASD refresh (aka upgrade) projects only added to the complexity. Pennies vs. pounds. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Staller, Allan allan.stal...@kbmg.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 10:04 AM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM- m...@listserv.ua.edu Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed)
On 2/22/2013 1:08 PM, Skip Robinson wrote: The biggest advantage of large volumes is reduction in UCB count. One disadvantage is that larger volumes take longer to back up and to restore. We have some astonishingly large 'volumes' in the open systems world that are supported by only modestly fast backup/restore processes. Keeping an important application down for a lengthy restore is a hard bullet to bite. Backup/restore times are definitely longer for large volumes. Having said that, our overall backup/dump times did not change after we consolidated many small volumes (mod3s, 6s, and 9s) onto fewer larger ones (mod27s, 54s, 81s, etc.). We need to a) every day incrementally backup 'n' changed tracks onto 'm' tape drives and b) periodically dump 'x' volumes onto 'y' tape drives where we stack multiple dumps onto high-capacity tapes. We have no VTS. No matter how we distribute the data among our DASD volumes (small, medium, large, extra-large), we never realize more than a few minutes of variability in overall backup or dump times. It seems that to appreciably reduce overall backup/dump time (without a radical hardware technology change) we can 1) purchase more tape drives, 2) update fewer DASD tracks every day (affects daily backups only), or 3) stop backing up or dumping some volumes. One area of focus is our corporate file server, which utilizes the DFS/SMB server on z/OS. We have large, single-volume ZFS data sets on large volumes. (They used to be multivolume ZFS on smaller volumes). One small update causes the entire ZFS to require daily HSM backup. That can be a time-consuming process. For example: ARC0722I BACKUP STARTING ON VOLUME LAN01 (SMS) AT 340 ARC0722I (CONT.) 02:36:48 ON 2013/02/23 SYSTEM MV60 ... ARC0723I BACKUP ENDING ON VOLUME LAN01 AT 04:37:58, 318 ARC0723I (CONT.) 0001 DATA SETS BACKED UP We do daily incremental backups of this data using Tivoli Storage Manager (The TSM 5.3 client runs under OMVS; the TSM 6.2 server runs on Linux for z). I'm tempted to stop backing up the physical ZFS altogether via HSM and rely 100% on the Tivoli backups. But, I'm hesitant to do so. We have never before excluded important volumes from our HSM backup/dump process. I'm curious what others would do in a similar situation... -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed)
On Sat, 23 Feb 2013 08:59:39 -0800, Ed Jaffe wrote: One area of focus is our corporate file server, which utilizes the DFS/SMB server on z/OS. We have large, single-volume ZFS data sets on large volumes. (They used to be multivolume ZFS on smaller volumes). One small update causes the entire ZFS to require daily HSM backup. That can be a time-consuming process. ... ITYM zFS. Most of my department's large UNIX filesystems are NFS mounted from ZFS (not zFS) on Solaris servers. Our daily backups are ZFS snapshots, almost negligible latency, followed by background dumps to tape. Disclaimer: I am associated with the Solaris/ZFS vendor, but not authorized to speak for them; my opinions are personal. -- gil -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Reducing Backup Time (Was: Mod-9 vs. Mod-27 vs. mixed)
On 2/23/2013 9:33 AM, Paul Gilmartin wrote: Most of my department's large UNIX filesystems are NFS mounted from ZFS (not zFS) on Solaris servers. Our daily backups are ZFS snapshots, almost negligible latency, followed by background dumps to tape. Right. FLASHCOPY of a large volume on our DS8100 can be done in almost an instant. But, copying the volume's data to tape takes hours -- as mentioned previously. My question is about whether a DFS/SMB ZFS should be backed up at all given its size and the existence of a daily TSM backup of its contents. -- Edward E Jaffe Phoenix Software International, Inc 831 Parkview Drive North El Segundo, CA 90245 http://www.phoenixsoftware.com/ -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
In the interest of frugality, I asked my storage guys some time ago to allocate some tiny volumes for JES checkpoint and couple data sets. After a while, they complained that it was more trouble than it was worth because we mirror most volumes to the DR site. For every tiny source volume, they needed a corresponding tiny mirror volume. Periodic DASD refresh (aka upgrade) projects only added to the complexity. Pennies vs. pounds. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Staller, Allan allan.stal...@kbmg.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 10:04 AM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM-MAIN@LISTSERV.UA.EDU Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
Hi, In our environment, we have a mix of emulated 3390-9 and 3390-54. When we have a need for small volumes (page, etc), we use 3390-9s and backfill them with rarely referenced files - documentation, etc. I tend to keep the system on 3390-9s (except SMPPTS!), as there was (once) a bit of a performance hit on the emulated 3390-54s. Not sure if that's still true. The physical drives are 650GB, and with the cache in front of them, I'm not sure it matters anymore. I'd be interested to hear opinions on this. Thanks! BobL -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Skip Robinson Sent: Friday, February 22, 2013 1:40 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Mod-9 vs. Mod-27 vs. mixed [ External ] In the interest of frugality, I asked my storage guys some time ago to allocate some tiny volumes for JES checkpoint and couple data sets. After a while, they complained that it was more trouble than it was worth because we mirror most volumes to the DR site. For every tiny source volume, they needed a corresponding tiny mirror volume. Periodic DASD refresh (aka upgrade) projects only added to the complexity. Pennies vs. pounds. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Staller, Allan allan.stal...@kbmg.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 10:04 AM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM-MAIN@LISTSERV.UA.EDU Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN This e-mail transmission may contain information that is proprietary, privileged and/or confidential and is intended exclusively for the person(s) to whom it is addressed. Any use, copying, retention or disclosure by any person other than the intended recipient or the intended recipient's designees is strictly prohibited. If you are not the intended recipient or their designee, please notify the sender immediately by return e-mail and delete all copies. OppenheimerFunds may, at its sole discretion, monitor, review, retain and/or disclose the content of all email communications. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
Someone once observed that 'rarely used' is not the same the thing as 'lightly used'. Whenever a file is accessed for read or write, IOS goes after it with a heavy boot. Then there's the problem of backups: if a file is worth keeping around, you probably want to back it up periodically, maybe even frequently. Files like JES checkpoint and couple data sets do not appreciate being made to stand in line even for seconds. The biggest advantage of large volumes is reduction in UCB count. One disadvantage is that larger volumes take longer to back up and to restore. We have some astonishingly large 'volumes' in the open systems world that are supported by only modestly fast backup/restore processes. Keeping an important application down for a lengthy restore is a hard bullet to bite. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Lester, Bob bles...@oppenheimerfunds.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 12:54 PM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM-MAIN@LISTSERV.UA.EDU Hi, In our environment, we have a mix of emulated 3390-9 and 3390-54. When we have a need for small volumes (page, etc), we use 3390-9s and backfill them with rarely referenced files - documentation, etc. I tend to keep the system on 3390-9s (except SMPPTS!), as there was (once) a bit of a performance hit on the emulated 3390-54s. Not sure if that's still true. The physical drives are 650GB, and with the cache in front of them, I'm not sure it matters anymore. I'd be interested to hear opinions on this. Thanks! BobL -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Skip Robinson Sent: Friday, February 22, 2013 1:40 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: Mod-9 vs. Mod-27 vs. mixed [ External ] In the interest of frugality, I asked my storage guys some time ago to allocate some tiny volumes for JES checkpoint and couple data sets. After a while, they complained that it was more trouble than it was worth because we mirror most volumes to the DR site. For every tiny source volume, they needed a corresponding tiny mirror volume. Periodic DASD refresh (aka upgrade) projects only added to the complexity. Pennies vs. pounds. . . JO.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 626-302-7535 Office 323-715-0595 Mobile jo.skip.robin...@sce.com From: Staller, Allan allan.stal...@kbmg.com To: IBM-MAIN@LISTSERV.UA.EDU, Date: 02/22/2013 10:04 AM Subject:Re: Mod-9 vs. Mod-27 vs. mixed Sent by:IBM Mainframe Discussion List IBM-MAIN@LISTSERV.UA.EDU Space wasted for small volumes e.g. XCF couple datasets. Just a talking point. With Hyper-PAV, etc. most of the other points are just hot air. The convenience of not having to support multiple geometries. snip A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. /snip -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
We are looking at converting about 12TB of 21TB of disk from Mod 9 to 27 (32,760). We are doing those storage groups that have 250GB of data in them so they use 10 M27s and adding another M27 adds 10% or less. Two storage groups have 1,600GB so adding 9 or 29.4 GB are very minor. We have some storage groups with a mix of M9s and M27s and we have not seen any problems. On Fri, Feb 22, 2013 at 11:48 AM, Tom Sims trs...@att.net wrote: A client with DS8000 DASD configured as a mix of 3390 Mod9 and Mod27s is considering a project to convert everything to Mod-27. Does anyone out there have some thoughts on the advantages or disadvantages of this? I'm not looking to start a religious discussion or other dinotribe, so feel free to respond to me privately. I'm just interested in perspective and talking points, either way. Thanks! Tom Sims zSystems Advocate -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
On Friday 22 February 2013 15:08:34 Skip Robinson wrote: Someone once observed that 'rarely used' is not the same the thing as 'lightly used'. Whenever a file is accessed for read or write, IOS goes after it with a heavy boot. Then there's the problem of backups: if a file is worth keeping around, you probably want to back it up periodically, maybe even frequently. Files like JES checkpoint and couple data sets do not appreciate being made to stand in line even for seconds. The biggest advantage of large volumes is reduction in UCB count. One disadvantage is that larger volumes take longer to back up and to restore. We have some astonishingly large 'volumes' in the open systems world that are supported by only modestly fast backup/restore processes. Keeping an important application down for a lengthy restore is a hard bullet to bite. . . JO.Skip Robinson If I remember correctly, there is also an issue of the number of paths available; fewer, larger volumes = fewer paths = more I/O contention. (This may be an old issue that is no longer relevant.) Leslie -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: Mod-9 vs. Mod-27 vs. mixed
If I remember correctly, there is also an issue of the number of paths available; fewer, larger volumes = fewer paths = more I/O contention. (This may be an old issue that is no longer relevant.) Pretty well with PAV HIPERPAV. - Ted MacNEIL eamacn...@yahoo.ca Twitter: @TedMacNEIL -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
