The restrictions covered here are those that are not covered elsewhere in
the documentation, or that are of a temporary nature and may be removed in a
later product. Permanent limits are generally covered in the main Help.
The notes below describe some known errors and restrictions in your
installation.
If you have installed a WrapPack containing updates and fixes to your
installation, the fixes and any last-minute issues are documented in the
WrapPack readme, which is downloadable with the WrapPack from the SupportLine
Web site. For details of how to get the latest information see
Updates and
SupportLine
- The following run-time features are not supported in 64-bit:
- HyHelp
- Enterprise Server. Some features might compile correctly, but
won't run.
- If you have configured Enterprise Server to support container-managed
services by defining external XA resource managers, any application-managed
service that issues CONNECTs to, or DISCONNECTs from, one of these defined
resource managers might receive warning- level response codes.
- The reference help for Mainframe Transaction Option includes topics
on some CICS transactions that are not supported and are not included in the
software. These are:
- A list of databases supported is missing from the section
Container- managed Services in the chapter
Introduction of your Enterprise Server Configuration and
Administration Guide. The list should follow the second paragraph
and should read:
- Under HP UNIX, PA-RISC, because of an issue with the operating
system, when a debuggable SEP is removed from the ES pool, it is not always
removed cleanly. This debuggable SEP might be 'terminated' and will show up in
console.log as 'terminating due to run-time error'. This should not affect your
ES system.
-536360
- If you upgraded from a previous version of Micro Focus Server and
have retained any enterprise servers that have a Web
Services listener, you will receive an error when you deploy
informing you that the listener cannot be found. To fix this, change the
listener name before deploying using one of the following methods:
- Dialog System does not run correctly on a terminal that has more
than 25 lines. To work around this set the LINES environment variable to 25.
- If the Web Browser tells you it's blocking active content, when you
view the Help, you need to allow active content. See the Software
Requirements section in
Hardware and Software
Requirements.
- You can view the Enterprise Server documentation in a separate
browser window while you are using the Enterprise Server Administration Web
interface, by clicking Contents in the menu at the left-hand
side of the page. For this link to work, you must have the following in place:
- An enterprise server that has a Web listener with the following
entry in the [virtual paths] section:
docs=<ES>/docs/html
- The enterprise server must be started
The default enterprise server, ESDEMO, may not have the correct
configuration information entry for its Web listener. If you encounter problems
accessing this documentation while you are using ESDEMO, you must correct the
entry using the Edit Listener page. Note that references to this configuration
entry in the context help and the Enterprise Server Configuration and
Administration Guide are no longer correct and you must use the
entry as specified above.
When you create a new enterprise server, a Web listener with the
required configuration information entry is automatically created.
- If the buttons on the Enterprise Server Administration Web page don't
work, you need to enable Active Scripting or JavaScript. See the Software
Requirements section in
Hardware and Software
Requirements.
-541326
- Fileshare does not correctly set offsets in 64-bit bytestream I/O
operations.
- The FSVIEW calls and commands that have to do with log files are not
currently implemented.
- On Red Hat Enterprise Linux 5.1 i686 (32 bit), the run-time system
error COBRT252 "Memory corruption detected (Fatal)”, might occur when running
deployed IMS service interfaces with tracing enabled. (To eanble tracing, you
check Generate Trace in the IMS pane of the Service Interface Settings dialog).
The workaround is to redeploy the service interface with the tracing flag
turned off.
-550132
- A CPU license for MTO or Micro Focus Server currently provides
unlimited use of the product. With this CPU license, if you use the license
manager command eslmfgetpv t to produce a license report, DEVMTO (or DEVES)
usage is not reported and 0 is shown against the appropriate license totals.
536449
- Threaded programs. If you create a multi-threaded application that
uses CBL_EXIT_PROC to perform an exit procedure when the program terminates,
and an error occurs during the exit procedure, the application does not
terminate with a non-zero exit code to the operating system. Shared object- C
module load failure.
- On IBM s/390, if you create a run-time system that includes a C
module and you use the run-time system to run a COBOL shared object that calls
a function within the C module, you will get an RTS 198 load error for that
function when "." (dot) is included as a directory to be searched on the load
library path (LD_LIBRARY_PATH).
- Deleting files. On HP-UNIX, if a program attempts to delete a file
that is already open to another user, the file status returned is 0/065 (File
locked), instead of the more usual 9/029 (Attempt to delete open file). This
applies to 32-bit files only: 64-bit files return the usual file status.
On some platforms the OpenLDAP client library is not supplied as
standard. As a consequence, if the Enterprise Server administrator adds a new
Security Manager that attempts to use the "mldap_esm" module supplied by Micro
Focus the following runtime errors may occur:
- If a Security Manager using "mldap_esm" is added to the Default ES
Security Manager List (the default is not to use an external Security Manager)
the following error will appear in the console when attempting to start an
Enterprise Server instance:
ESFMI0201C Unable to load ESM Module "[Security Manager name]"
(mldap_esm)
Similarly, if such a Security Manager is added to the security
configuration for a specific Enterprise Server instance, over-riding the
default ES security, the same error will occur.
- If a Security Manager using "mldap_esm" is added to the Directory
Server Security Manager List (the default is to use the "MFDS Internal
Security" Manager), the following error will be displayed:
SF0024E Error 16/4/0 SAFMGR initialization failed:
Failure loading ESM Module
The Security Manager will not be used by Directory Server
Security.
- If OpenLDAP client support is not present, MF Directory Server
repository location configuration will also display an error if the MF
Directory Server repository location is changed to an LDAP server partition:
MDS1503E Schema check error: 501, OTHER.
Previous schema path restored.
Some UNIX systems may already have OpenLDAP installed. On other
systems, you may wish to use the OpenLDAP provided by Micro Focus, if it is
included with the product on your platform. See the
$COBDIR/bin/configure_openldap script and
$COBDIR/etc/readme-ldap.txt for more information.
- An Enterprise Server that is configured to use a Security Manager
that uses the mldap_esm module (that is, using LDAP for security) cannot be
started using a file repository, either by using the -m option to casstart and
specifying a file path, or by running casstart when MFDS is not available.
-542314
- A new facility called passtokens can be configured to let
administrators move back and forth between MFDS and ESMAC without signing in
each time they switch from one to the other. (This requires that MFDS and
Enterprise Server use the same security configuration, and that the user has
appropriate privileges.) However, the passtoken is created when the HTML page
with the link to the other facility (for example, the MFDS "server details"
page with the ESMAC link) is generated, and passtokens expire after a short
time. If you wait too long before clicking the link to go from MFDS to ESMAC or
from ESMAC to MFDS, you will be forced to sign on again.
- 543176
- Customers who use the audit feature of Enterprise Server security
should be aware that we cannot enforce auditing for administrative changes to
the security configuration (adding or removing users, changing permissions,
etc). A user with sufficient privilege (an administrator) can disable auditing
before performing other administrative changes, or use third-party tools to
update the configuration in the ESM without using any Micro Focus tooling at
all.
-543360
- When external security is enabled for Enterprise Server, user
credentials (username and password) have to be supplied to casstart in order to
start the region. casstart uses those to do an ESF Verify and Auth to ensure
that the user has authority to start the region. casstart also uses those
credentials to bind to MFDS.
A problem can arise because MFDS is a different security domain from
Enterprise Server and may have a different security configuration, particularly
if Enterprise Server is using an ESM (such as the eTrust one) that doesn't
support MFDS. Consequently, administrators may have to define user accounts
twice, once in the ESM used by Enterprise Server, and once in MFDS internal
security (or a different ESM used by MFDS), in order to let those users start a
region.
-543472
- If you have clients using the ECI interface with the CICS
Transaction Gateway to communicate with a CICS application running in Mainframe
Transaction Option (MTO), a client program might receive an
ECI_ERR_TRANSACTION_ABEND return code with an abend code of AZTY, indicating
that several clients tried to communicate with the system at the same time but
not all requests could be satisfied by MTO. The workaround for this issue is to
set the environment variable TX_UNICLI_SESSIONS to the required number of
sessions in the Configuration Information for the enterprise server as follows:
[ES-Environment]
TX_UNICLI_SESSIONS=<required-number-of-sessions>
The default is 4.
You should set "Maximum requests" in the Client Section of the
CICS Transaction Gateway configuration to the same value.
If you do this you will also probably need to change the client
program to ignore the ECI-ERROR_MAX_SESSIONS return code and retry the call.
For example
do
{
Rc = CICS_ExternalCall (&EciParms);
} while (Rc == ECI_ERR_MAX_SESSIONS);
If you do not make this modification, the client program might
still receive the ECI_ERR_TRANSACTION_ABEND return code.
- SORT/MERGE using old .int files. A program error might occur if you
are executing intermediate code files that were created using previous versions
of COBOL for UNIX and which include SORT/MERGE statements.
- Sorts using the NLS directive. Programs compiled with the NLS
directive that make use of COBOL SORT with alphanumeric keys return records in
the standard alphanumeric order, rather than in the correct order for the NLS
locale. This restriction only applies to file sorts and not to table sorts.
- 514543
- When a COBOL service that is processing a transactional J2EE client
request using the local connector (mfcobol-localtx.rar) fails by ending with a
fail return-code, Enterprise Server displays the message:
Cannot end/commit/rollback/recover a transaction that doesn't exist
This message is incorrect, as in this case rollback has succeeded
and your data is in a stable state.
-533220
- DISPLAY-OF intrinsic function. When using the DISPLAY-OF intrinsic
function to convert Unicode characters to your current locale, any Unicode
characters that have no equivalent in the current locale might not be replaced
by the optional default replacement character on some UNIX platforms.
-532366
- ICONV_BYTEORDER. On the TRU64 platform, in order for the Unicode
support to work, you need to set the ICONV_BYTEORDER environment variable as
follows: ICONV_BYTEORDER=little-endian export ICONV_BYTEORDER.
With stateful EJB generated for EJB 3.0, there is a problem using the
test client. All test clients use the same EJB instance. There is also an issue
with stateful EJB timeout. There is no way to set timeout programmatically for
EJB 3.0. This is a WebSphere limitation at this time.
- 556818
- If you install the 5.1 WrapPack 2 release over 5.1 GA or 5.1 WrapPack
1 - and perform a subsequent uninstall of WrapPack 2- you will need to manually
restart MFDS to ensure services are restarted correctly.