SFS Abend

[Schuh, Richard]

We just tried using a Dataspace enabled DIRC directory for the first time and ran into a bit of a problem. The directory is large - 546000 blocks committed to only 201 files. The owner started 12 machines that all access this directory at the same time. The Filepool server crashed with a 0C5 abend. Unfortunately, the server is running on a back-level CMS, CMS16, so I don't want to fix that code. Two questions:

1. Is this a problem that has likely been fixed in a later release? We are planning to go to z/VM 5.2 and CMS22 in about a month. It would be possible to switch the FP server to CMS22 early provided that all clients did not also have to be switched.
2. Would having 1 user access the directory early to [prepare the DS do anything to help? It would at first blush appear that there might have been a clash between the other eleven machines and the FP server, which was probably still trying to create the DS that they wanted to access. At my previous shop, we did have DS eligible DIRCs, including one that was large, both in number of files and committed space. We did prime the pump for the large directory by having a service machine access it during start-up. It was accessed and read by 100s of users every day. We never had a problem with it residing in a DS enabled DIRC.

Regards,
Richard Schuh

[Hodge, Robert L]

Richard,
A search string of "SFS ABEND0C5 LVLS/310" in IBMLink found APAR
VM63105, "ABEND0C5 WHEN PLACING SFS DIRECTORY IN DATASPACE". The PTF for
CMS 16 is UM30412.
1) Yes, it is fix in a later release. And it is fixed in CMS 16, if you
install the PTF.

-----Original Message-----
From: VM/ESA and z/VM Discussions [mailto:VMESA-***@LISTSERV.UARK.EDU] On
Behalf Of Schuh, Richard
Sent: Friday, February 17, 2006 2:07 PM
To: VMESA-***@LISTSERV.UARK.EDU
Subject: SFS Abend

We just tried using a Dataspace enabled DIRC directory for the first
time and ran into a bit of a problem. The directory is large - 546000
blocks committed to only 201 files. The owner started 12 machines that
all access this directory at the same time. The Filepool server crashed
with a 0C5 abend. Unfortunately, the server is running on a back-level
CMS, CMS16, so I don't want to fix that code. Two questions:

1. Is this a problem that has likely been fixed in a later release?
We are planning to go to z/VM 5.2 and CMS22 in about a month. It would
be possible to switch the FP server to CMS22 early provided that all
clients did not also have to be switched.
2. Would having 1 user access the directory early to [prepare the
DS do anything to help? It would at first blush appear that there might
have been a clash between the other eleven machines and the FP server,
which was probably still trying to create the DS that they wanted to
access. At my previous shop, we did have DS eligible DIRCs, including
one that was large, both in number of files and committed space. We did
prime the pump for the large directory by having a service machine
access it during start-up. It was accessed and read by 100s of users
every day. We never had a problem with it residing in a DS enabled DIRC.

Regards,
Richard Schuh