These answers to technically oriented Frequently Asked Questions about FLEX-ESTM should help to familiarize you with our product. For further information, please contact Fundamental Software.
This FAQ addresses technical aspects of FLEX-ES only. For general and business related matters, please see our General FAQ.
Question: What kinds of tape support does FLEX-ES provide?
Q: Can FLEX-ES channel-attached 3480s operate at 4.5 megabytes/second?
A: Yes. The FLEX-ES Parallel Channel Adapter (PCA) supports 3480/3490 series (and compatible) tape drives in Data Streaming mode at the full speed of up to 4.5 megabytes/second.
Q: How does the FLEX-ES FakeTapeTM virtual tape technology fit in to conventional installations?
A: FakeTape virtual tapes are advantageous whenever the following are required:
FakeTapes are fast, and they require less activity on the part of an operator (they may be mounted by typing a command, rather than by physically mounting a tape). They are therefore appropriate, for example, as scratch tapes.
When these conditions are met, the operational savings possible with FakeTape can be significant. One FLEX-ES customer was able to reduce physical tape mounts from 15,000 per month to 150 per month by replacing most physical tapes with FakeTapes.
FLEX-ES FakeTape emulated tape drives appear to the mainframe operating system as if they were 3420 series mainframe tape drives. No special mainframe OS support is required for their use. They read and write emulated "tapes" from and to disk files on the host UNIX server. These FakeTape files may be "mounted" and "unmounted" from the FLEX-ES FakeTape drives by using commands at the FLEX-ES console (the Command Line Interface, or CLI).
For example, if a FakeTape emulated tape exists on the host UNIX server's disks and is named, say, "/tapes/tape2", this tape may be mounted on a FLEX-ES FakeTape drive at mainframe device address 180 by issuing the following command at a FLEX-ES CLI console:
mount 180 /tapes/tape2
The "mount" command is equivalent to physically mounting a conventional tape and pressing the "LOAD" button.
Automating FakeTape Operations
The FLEX-ES console (CLI) system is programmable from the UNIX host using UNIX shell scripts and other UNIX programming languages. This allows tape operator commands such as mount to be automated.
Because FakeTape emulated "tapes" are UNIX disk files, they cannot be exchanged with non FLEX-ES installations which expect physical tape media. It is possible, however, to use a standalone utility provided with FLEX-ES to move FakeTape files to SCSI-attached 9-track tape.
However, FakeTape tapes may be exchanged with other FLEX-ES systems.
The FLEX-ES "network channel" capability allows emulated devices on any FLEX-ES host in a TCP/IP network to be attached to FLEX-ES CPUs on any other host in the network, regardless of the physical location of the two hosts. It is thus possible to write FLEX-ES FakeTape tapes to hosts remote from the FLEX-ES CPU's host; the FakeTape drive on the remote host would appear to be a local device channel-attached to the FLEX-ES CPU's host. (The performance of the network must of course be taken into consideration.) This presents possibilities for offsite backup and for disaster recovery.
FSI provides a FakeTape Programming Library which has interfaces and routines for creating, reading, and writing FakeTapes from native UNIX applications on any FLEX-ES UNIX host. This means that, with additional programming, FakeTapes can be used as a medium for exchanging data between native mainframe applications and native UNIX applications.
Q: Are ECKD emulated DASD supported?
Q: How much emulated DASD can I have?
A: The number of emulated FLEX-ES DASD present on a system is limited by:
Small UNIX servers are now typically configured with tens of gigabytes of disk. Medium sized servers may be configured with hundreds of gigabytes. Large servers can now accommodate a terabyte or more of disk. These large server disk arrays can accommodate very many FLEX-ES emulated DASD.
Q: Can I mix 3390 and 3380 emulated DASD?
A: Yes. All types of FLEX-ES emulated CKD and FBA DASD can be used, simultaneously, on a single FLEX-ES.
Q: If a FLEX-ES emulated DASD fails, does the mainframe operating system get notified?
A: Yes. In the case of a complete failure of a disk or RAID array on the host UNIX server which causes the failure of one or more FLEX-ES emulated DASD, the FLEX-ES mainframe and the mainframe operating system running on it are notified of this failure by sense data returned from the emulated DASD.
If one host UNIX server disk fails in a redundant (RAID) disk array, then the disk array will correct that failure in a manner transparent to both FLEX-ES and any mainframe OS running on top of FLEX-ES. From the point of view of FLEX-ES and the mainframe OS, the disk will not have failed.
Q: Where are FLEX-ES emulated DASD cached?
A: In RAM on the host server (not in emulated mainframe main storage). DASD caches are transparent to the activity of the emulated mainframe. FLEX-ES emulated DASD are implemented using:
Write-Through or Write-Back
Emulated DASD caches may be configured either as write-through or write-back caches, and may be set up to be any size above certain default minimums (depending on the RAM available on the host).
DASD Cache Details
Emulated DASD caching is implemented on a per-control unit basis as described below:
DASD Cache Size Definition
Each emulated DASD attached to a given emulated control unit may be defined to have a minimum amount of cache allocated for it at the emulated control unit. This amount is specified in terms of whole cylinders, where the cylinder size depends on the particular emulated DASD type. If no minimum is specified, the default is one cylinder. For example, a cylinder on a 3380 has 15 tracks, so the default minimum cache allocated for a 3380 is one cylinder, or 15 3380 tracks. Larger minimum values may be defined.
Control Unit Cache Size Definition
Each emulated control unit implements at least as much cache as the total sum of the cache sizes specified for all of the DASD attached to that control unit. Thus, if an emulated DASD control unit has 4 emulated 3380 DASD attached to it, and each has the default minimum of 15 tracks (1 cylinder each) of cache, and the control unit had no control unit minimum defined, then the control unit will allocate 60 tracks of cache (15 + 15 + 15 + 15). If one of the emulated DASD had, for example, a defined minimum of 30 tracks, then the control unit would define 75 tracks of cache (30 + 15 + 15 + 15).
Floating Cache Capacity
Control units may have more cache defined than the minimum required for all of their attached DASD. For example, a control unit which attaches 4 emulated 3380s, each with the default minimum of 15 tracks of cache, may be defined to have not 60 (the minimum) but 120 tracks of cache. If extra cache capacity such as this is defined, then this extra capacity "floats" between the DASD on the control unit and is used by the DASD which need it most. This is an extremely efficient use of server RAM to provide cache resources.
Q: Does the FSI PCA (Parallel Channel Adapter) support the attachment of DASD?
Q: What is the relationship between VM minidisk I/O and FLEX-ES I/O?
A: They are independent. FLEX-ES provides an emulated mainframe. A mainframe operating system such as VM runs on it just as it would on a conventional physical mainframe.
Q: Does FLEX-ES handle expanded storage?
Q: What is the relationship between the VM, MVS, or VSE I/O Configuration Program (IOCP) and FLEX-ES I/O configuration?
A: IOCP is an IBM product which supports the configuration of channels or channel paths, control units, and I/O devices in an IBM mainframe environment. FLEX-ES does not support IOCP but provides equivalent functionality using other methods.
FLEX-ES system (CPU), channel or channel path, and device configuration is defined by configuration files and maintained by a FLEX-ES Resource Manager.
Copyright © 1998-2012 by
Fundamental Software, Inc.
Document Version 7.1