IBM Virtual Tape TS7720 OUT OF CACHE; BIM-EPIC is the tape management system

[Sherry Everhart]

Hello Everyone,

We are currently unable to process any virtual tape requests and have been down for 24-hours with this issue. Is there anyone out there who can help me?

We run BIM-EPIC 7.3B under z/VSE4.3 and utilize an IBM TS7720 Virtual Tape System. Yesterday we experienced some errors in our virtual tape mounts/dismounts in our VSE jobs so I logged on to the TS7720 and found two error messages: HYDME1200W - One or more clusters are operating with limited free space in cache, and HYDME1201W - One or more clusters have run out of cache resources.
We only have one cluster and there are no internal tape drives on our TS7720.

I've spoken with IBM and they directed me to BIM-EPIC. I've spoken with BIM-EPIC and they directed me back to IBM. IBM said it's not a hardware issue and that the CACHE is managed by the tape management system (i.e. EPIC). BIM-EPIC said that their software is functioning as it should so I need to talk to IBM about what's wrong.

I want to know what is stored in the TS7720 CACHE and how do I remove the older data, as the action (below) states I should?

Thanks for the help.

Sherry


The HYDME1201W error message states:
-----------------

A disk-only TS7720 Virtualization Engine does not attach to a physical backend library; all virtual volumes are stored in the cache. When the allocated free space is less than 1 TB level, the cluster enters the Out of Cache Resources state. Mount operations that have been accepted prior to entering this state will complete and volumes currently mounted can continue to perform any host I/O operation. Any fast-ready mount operations received while in this state fail with completion and reason codes indicating that the mount could not be processed due to insufficient cache space. Specific volume mount operations received while in this state are allowed, however, any attempt to write to the volume fails with a write-protect error. The cluster exits this state when the allocated free space is more than 3.5 TB. If the cluster is in a grid configuration, the Out of Cache Resources state reflects the composite state of cache in all TS7720 Clusters in the grid.


Action

1.Check the status of any degraded cluster on the "Cluster Summary" page.
2.Remove older data from the TS7720 Cache until the allocated free space is more than 3.5 TB.
--------------------

[Kevin Corkery]

My best guess would be that you need to purge some unneeded tape images from the device.  Interesting the cutoff being 1TB to place you into this state and the 3.5TB needed to get out.  Sounds very "z/OS scale" to me.  My virtual tape has only 1.5TB in total and I'm only using about 1TB actively. 
 

> To: vs...@lehigh.edu
> Date: Fri, 19 Aug 2016 04:03:12 -0700
> From: seve...@maccnet.com
> Subject: IBM Virtual Tape TS7720 OUT OF CACHE; BIM-EPIC is the tape management system
>
> Hello Everyone,
>
> We are currently unable to process any virtual tape requests and have been =
> down for 24-hours with this issue. Is there anyone out there who can help =
> me?
>
> We run BIM-EPIC 7.3B under z/VSE4.3 and utilize an IBM TS7720 Virtual Tape =
> System. Yesterday we experienced some errors in our virtual tape mounts/di=
> smounts in our VSE jobs so I logged on to the TS7720 and found two error me=
> ssages: HYDME1200W - One or more clusters are operating with limited free =
> space in cache, and HYDME1201W - One or more clusters have run out of cache=
> resources.
> We only have one cluster and there are no internal tape drives on our TS772=
> 0.
>
> I've spoken with IBM and they directed me to BIM-EPIC. I've spoken with BI=
> M-EPIC and they directed me back to IBM. IBM said it's not a hardware issu=
> e and that the CACHE is managed by the tape management system (i.e. EPIC). =
> BIM-EPIC said that their software is functioning as it should so I need to=

> talk to IBM about what's wrong.
>

> I want to know what is stored in the TS7720 CACHE and how do I remove the o=

> lder data, as the action (below) states I should?
>
> Thanks for the help.
>
> Sherry
>
>
> The HYDME1201W error message states:
> -----------------
>

> A disk-only TS7720 Virtualization Engine does not attach to a physical back=
> end library; all virtual volumes are stored in the cache. When the allocate=
> d free space is less than 1 TB level, the cluster enters the Out of Cache R=
> esources state. Mount operations that have been accepted prior to entering =
> this state will complete and volumes currently mounted can continue to perf=
> orm any host I/O operation. Any fast-ready mount operations received while =
> in this state fail with completion and reason codes indicating that the mou=
> nt could not be processed due to insufficient cache space. Specific volume =
> mount operations received while in this state are allowed, however, any att=
> empt to write to the volume fails with a write-protect error. The cluster e=
> xits this state when the allocated free space is more than 3.5 TB. If the c=
> luster is in a grid configuration, the Out of Cache Resources state reflect=

> s the composite state of cache in all TS7720 Clusters in the grid.
>
>
> Action
>
> 1.Check the status of any degraded cluster on the "Cluster Summary" page.

> 2.Remove older data from the TS7720 Cache until the allocated free space is=

> more than 3.5 TB.
> --------------------
>

> _______________________________________________
> VSE-L mailing list
> VS...@lists.lehigh.edu
> https://lists.lehigh.edu/mailman/listinfo/vse-l

[Sherry Everhart]

Our issue is now resolved. Thanks, Kevin, for your reply.

Sherry

[Ken Bloom]

What was the resolution?

Sent from my iPhone

> On Aug 21, 2016, at 5:13 AM, Sherry Everhart <seve...@maccnet.com> wrote:
>
> Our issue is now resolved. Thanks, Kevin, for your reply.
>
> Sherry
>

[Kevin Corkery]

So, what was the resolution?