APAR: IY13310 COMPID: 5639I3500 REL: 310
ABSTRACT: EFFECTIVE (STRICTEST) PWDMINLEN POLICY NOT ENFORCED

PROBLEM DESCRIPTION:
When assigning or changing a user's password, the strictest
policy (default registry vs. user's primary org policy) is
not adhered to, thus allowing a user to enter a password
shorter in length than should be allowed.

------

APAR: IY13619 COMPID: 5639I3900 REL: 310
ABSTRACT: DTSD REQUESTS SHM SEG SMALLER THAN OS MINIMUM, THEN DIES

PROBLEM DESCRIPTION:
 DTSD requests shm seg smaller than OS minimum, then dies

LOCAL FIX:
Get the system default for the minimum and if the requested
is less than that bump it to the minimum.

------

APAR: IY15261 COMPID: 5639I3500 REL: 310
ABSTRACT: SECD CORE DUMP IN RS_ATTR_UPDATE

PROBLEM DESCRIPTION:
secd core dump in rs_attr_update() from Liberty Mutual.

LOCAL FIX:
We're getting a core dump because earlier in the routine
rs_attr_update_verify_under_RIW() returned a bad status
(0x17122143, schema entry not found) and therefore
existing_attrs had never been filled in and aud_lkup_num_attrs
was uninitialized. So we were trying to free uninitialized
data structures.
The fix: Initialize aud_lkup_num_attrs to 0 and aud_lkup_status
to a non-zero value.

------

APAR: IY15562 COMPID: 5639I3500 REL: 310
ABSTRACT: DFSBIND SPINS AND CONSUMES EXCESSIVE CPU TIME

PROBLEM DESCRIPTION:
dfsbind appears to be hung, spinning in a loop which
consumes excessive CPU time
At some point dfsbind goes into a spin and starts using
excessive CPU time. If one attaches via dbx to the dfsbind
process, the stack should look similar to:
(dbx) th current 6;t
krb5_fcc_open_file(??, ??) at 0xd04c6cbc
krb5_fcc_next_cred_aux(??, ??, ??, ??) at 0xd05c09e8
krb5_fcc_store(??, ??) at 0xd05c31c4
krb5_get_credentials_dce(??, ??, ??, ??) at 0xd05a3a84
sec_krb_get_cred(??, ??, ??, ??, ??, ??, ??, ??) at 0xd05a1da8
cred_from_cred_rep(??, ??, ??, ??) at 0x100060b4
handle_sec_krb_get_cred(??, ??, ??) at 0x10004700
do_auth_request(??, ??, ??, ??) at 0x10002e1c
ProcessRequest(??) at 0x100021f4
service_thread(??) at 0x10001b58
pthread._pthread_body(??) at 0xd012c358

LOCAL FIX:
Stop and restart dfsbind.

------

APAR: IY15602 COMPID: 5639I3500 REL: 310
ABSTRACT: UPGRADE_INSTALL TIMEOUT TOO SHORT

PROBLEM DESCRIPTION:
During a upgrade_install the timeout for secd
may be too short when there is a large security
database to convert.

------

APAR: IY15610 COMPID: 5765E5400 REL: 440
ABSTRACT: DIFFERENT BEHAVIOR ON 4.4.0 FOR DARE MIGRATED RG WITH STICKY

PROBLEM DESCRIPTION:
In HA 4.4.0 if a rotating resource group is DARE migrated with
the "sticky" flag and then the "sticky" node is brought down
with takeover and then brought back up. The rotating resource
group stays on current location, it should fall-back to "sticky"
location as it does in HA 4.3.1.

PROBLEM SUMMARY:
a rotating resource group which had been dare migrated with
the sticky flag, followed by a node down on the sticky node
will not fall-back to the sticky node when the sticky node
is re-integrated, but will instead follow standard rotating
behavior of remaining on current node.

PROBLEM CONCLUSION:
modify algorithm to allow fallback behavior in rotating groups
in the case of a sticky node

------

APAR: IY15640 COMPID: 5639I3500 REL: 310
ABSTRACT: RMXCRED CORE DUMPS IN SEC_LOGIN_PVT_FREE_LOGIN_CONTEXT

PROBLEM DESCRIPTION:
rmxcred core dumps with the following stack trace:
Segmentation fault in dce_free at 0xd042b2c0 ($t1)
0xd042b2c0 (dce_free+0x34) 8003ffd8 l r0,-40(r3)
(dbx) where
dce_free(??) at 0xd042b2c0
libdcelibc_r.free(??) at 0xd07a9488
sec_login_pvt_free_login_context(??) at 0xd04c88e0
sec_login_pvt_allocate_login_context() at 0xd04ce428
sec_login_pvt_get_default_context(??) at 0xd04cec88
sec_login_get_current_context(??, ??) at 0xd04bde80
seek_and_destroy(??) at 0x10001288
main(??, ??) at 0x10002120

------

APAR: IY15698 COMPID: 5639I3900 REL: 310
ABSTRACT: SOLARIS CONFIG REQUIRES WHOAMI WHICH IS NOT PRESENT IN DEFAULT

PROBLEM DESCRIPTION:
When configuring DCE 3.1 on Solaris 2.7 config.dce requires
/usr/ucb/whoami which is not installed by default, but is
part of Sun Package SUNWscpu. Until this package is installed
this configuration will not work.

------

APAR: IY15820 COMPID: 5639I3500 REL: 310
ABSTRACT: SECD CORE DUMPS IN RSDB_ATTR_LOOKUP_BY_NAME

PROBLEM DESCRIPTION:
The master secd core dumps with the following stack trace
during an "account create" operation:
(dbx) t
signal.pthread_kill() at 0xd03c9f80
signal._p_raise() at 0xd03c946c
raise.raise() at 0xd017510c
abort() at 0xd016e860
dce_svc_printf_withfile() at 0xd04b2788
rsdb_attr_lookup_by_name() at 0x100682f4
rs_pwd_mgmt_strength_check() at 0x10111f18
rs_acct_add() at 0x10123498
krb5rpc_sstub.ssrs_common_sstub() at 0x1014d808
rpc_ssr_sstub() at 0xd04ec580
rpc__cn_call_executor() at 0xd0682a94
cthread_call_executor() at 0xd04e4ff8
pthread._pthread_body() at 0xd03bd358
In addition, the following message will be written to the
fatal.log:
dce_err: 0x17122eb9: unable to retrieve attr_list (dce / sec)
This problem is the result of incorrect usage of locks
surrounding access (r/w) to the rgy database, specifically,
the attribute database. The 2 instances of this problem
seen thus far were 3 months apart, during the course of
running a script which added principals and accounts
non-stop in the thousands. Therefore, it is not considered
to be a very prevalent problem.
The basic scenario:
-- We're attempting to add a new account.
-- We make a call to rs_pwd_mgmt_strength_check, but before that
   call we drop the lock on the db in order to prevent a
   deadlock should we have to do an rpc out to the password
   management server.
-- However, within rs_pwd_mgmt_strength_check we do some more
   db lookups (namely on the attribute list of the principal)
   which are now not protected by a read lock.
-- Apparently another thread had obtained a write lock on the
   db and was in the middle of making an update.
-- This causes the rsdb_attribute_lookup_by_name to attempt to
   read the db in an unknown state (i.e. in the middle of a
   write) and thus we get the unpredictable result of "attrlist
   not found".

------

APAR: IY15856 COMPID: 5639I3700 REL: 310
ABSTRACT: PE SITE FILE DOESN'T GET UPDATED WITH DCED -T 0 OPTION

PROBLEM DESCRIPTION:
The problem is when dced is configured in a client machine
with -t 0 option, the pe_site file has only the master server
and other servers in cell is not there. chpesite command
also does not update the file.

LOCAL FIX:
If -t 0 option is mentioned with dced then, the pe_site file
will be built with ALL the servers and after that automatic
update will NOT be done. But chpesite command will still be
able to update the pe_site file.

------

APAR: IY15863 COMPID: 5639I3700 REL: 310
ABSTRACT: CDS PERFORMANCE PROBLEM WHEN THE MASTER CLH IS DOWN

PROBLEM DESCRIPTION:
A significant performance degradation in CDS
lookups when the master cds server is down.

LOCAL FIX:
Changes were made to CACleSetOk() function in
src/directory/cds/cache/cacle.c and in
src/directory/cds/library/cds_check_ch.c for GDA.

------

APAR: IY15892 COMPID: 5765E5400 REL: 440
ABSTRACT: CONCURRENT VARYONVG FAILS AFTER APAR IY14247 INSTALLED - HACMP,H

PROBLEM DESCRIPTION:
The customer was running in concurrent mode with a fibre
attached 2105f20 and after applying an AIX PTF which pulled
in HACMP PTF's they could not bring up the cluster because
their concurrent vg's failed to varyon.

PROBLEM SUMMARY:
****************************************************************
* USERS AFFECTED:
* Systems using concurrent RAID devices running fileset level
* cluster.base.server.events.4.4.0.4 or
* cluster.es.server.events.4.4.0.4
****************************************************************
* PROBLEM DESCRIPTION:
* Unable to start cluster services after applying 4.4.0.4 level
* of cluster.base.server.events or cluster.es.server.events
* on systems using a concurrent RAID device due to a failure
* to varyon the concurrent volumegroups.
****************************************************************
* RECOMMENDATION:
* Order APAR IY15892.
****************************************************************

PROBLEM CONCLUSION:
Corrected the logic in cl_mode3 for checking whether the
vg was defined as concurrent capable or not, and if
concurrent capable, whether the old or new concurrent
capable. Prior to this change it was not correctly recognizing
old definition for RAID devices, so did not use convaryonvg as
it should have.

------

APAR: IY16047 COMPID: 5765E5400 REL: 440
ABSTRACT: CLUSTER VERIFY OR SYNCH FAILS WITH MSG INDICATING CLUSTER.LOG FI

PROBLEM DESCRIPTION:
When the customer tried to run cluster verification or synch
it would fail with the only message indicated:
cllog: The cluster.log log file has already been redirected via
modification of the /etc/syslog.conf file on node <nodename>.
If you wish to redirect this log again, please change the
cluster.log entries of the /etc/syslog.conf file. Inspection
of the /etc/syslog.conf file on the cluster nodes, however, did
not show any apparent problem.

LOCAL FIX:
Remove tabs between fields of /etc/syslog.conf file entries.

PROBLEM SUMMARY:
When the customer tried to run cluster verification or synch
it would fail with the only message indicated:
cllog: The cluster.log log file has already been redirected via
modification of the /etc/syslog.conf file on node <nodename>.
If you wish to redirect this log again, please change the
cluster.log entries of the /etc/syslog.conf file. Inspection
of the /etc/syslog.conf file on the cluster nodes, however, did
not show any apparent problem.

PROBLEM CONCLUSION:
The entries in the syslog.conf file for cluster.log had
tab(s) instead of space(s) between fields, and the code that
was parsing these entries was using cut with space as
field separator, so did not separate the fields. The code
was changed to use awk rather than cut in cllog.sh,
clsnapshot.sh, and cld_logfiles.sh.

------

APAR: IY16332 COMPID: 5639I3700 REL: 310
ABSTRACT: OLD AUDIT DIRECTORY PREVENTS SECD FROM STARTING

PROBLEM DESCRIPTION:
secd (or possibly any other dce daemon using audit) does not
start with the following message found in the fatal.log:
2001-02-01-14:46:30.256-06:00I----- secd FATAL sec rs_mainthr
rs_main.c 1526 0x3 005b290 msgID=0x10323078 Cannot make
directory
This situation arises when a previously running dce daemon
(using audit) apparently does not shutdown gracefully, allowing
for proper cleanup of its /var/dce/audit/client/<pid>
directory. When using audit, subdirectories are created in the
afore-mentioned directory using the pid as the directory name.
If a new dce daemon is starting with a pid equal to an
already-existing audit subdirectory, then it will fail to
start.

LOCAL FIX:
Simply trying to restart the dce daemon again will likely be
successful since the odds of obtaining the exact pid as before
are extremely low. Or to be absolutely certain, one could get
a "ps" listing of the currently running dce daemons and
determine which <pid> subdirectory in /var/dce/audit/client is
no longer in use and remove that directory and its contents.

------

APAR: IY16359 COMPID: 5639I3700 REL: 310
ABSTRACT: MAX OBJECT NAME LENGTH IN CDS CAUSES ERROR CDS_NOROOM

PROBLEM DESCRIPTION:
If an object is created in CDS, the defined maximum name
length results in the error CDS_NOROOM, which means the
entered name is too long.

PROBLEM SUMMARY:
Creating a cds full name with length greater than 402
characters will return with the error CDS_NOROOM.

PROBLEM CONCLUSION:
Changes are made to the CDS code so that there is a
consistency between the documented full name length and the
length supported in the CDS code. There is a further
inconsistency in the CDS code itself, where sometimes
it process the max length as 402 and at other times, as 1023.

------

APAR: IY16440 COMPID: 5765E5400 REL: 440
ABSTRACT: CLVER -R DISPLAYS NO /.RHOST MESSAGE IF LANG=EN_US

PROBLEM DESCRIPTION:
If the message catalog is used clverify of cluster resources
will fail with following messages.
/.rhosts file does not exist on node .
On network , the service address associated with
(Memory fault(coredump)
To work around the problem set LANG=C
        export LANG=C
prior to running clverify.

PROBLEM SUMMARY:
If the message catalog is used clverify of cluster resources
will fail with following messages.
/.rhosts file does not exist on node .
On network , the service address associated with
(Memory fault(coredump)

PROBLEM CONCLUSION:
Change clver.h CLVER_CUSTOM_DISKINFO 155 through 158 to 955
through 958.

------

APAR: IY16510 COMPID: 5765E5400 REL: 440
ABSTRACT: INTERFACE ROUTE LOST AFTER SWAP_ADAPTER EVENT - HACMP,HAES

PROBLEM DESCRIPTION:
Evidently, after upgrading to a certain level of AIX 4.3.3,
after a swap_adapter event occurred there was no route on the
interface to which the service address had been swapped. This
route should normally be created as a result of the ifconfig
of the interface to the up state by the HACMP script, since
the previous routes were deleted.

PROBLEM SUMMARY:
Evidently, after upgrading to a certain level of AIX 4.3.3,
after a swap_adapter event occurred there was no route on the
interface to which the service address had been swapped. This
route should normally be created as a result of the ifconfig
of the interface to the up state by the HACMP script, since
the previous routes were deleted.

PROBLEM CONCLUSION:
Though the HACMP script has been working across several
different versions and levels of AIX and HACMP through
use of deleting the route and ifconfig down of the interfaces,
at some level of change within AIX 4.3.3 this is no longer
consistent. As recommended by AIX TCP I have changed the
cl_swap_IP_address script to detach both interfaces before
ifconfig either of them back up.

------

APAR: IY16663 COMPID: 5765E5400 REL: 440
ABSTRACT: PATH FOR THE GREP CMD IS INCORRECT IN THE SCRIPT, SSA_CONFIGURE

PROBLEM DESCRIPTION:
Path for the grep cmd is incorrect in ssa_configure script.
The check_err() routine has a line bin/grep without the
leading /.

PROBLEM CONCLUSION:
Add missing leading /.

------

APAR: IY16727 COMPID: 5765E5400 REL: 440
ABSTRACT: C-SPOC LVM FUNCTIONS FAIL WHEN GHOST DISKS EXIST - HACMP,HAES

PROBLEM DESCRIPTION:
The customer had booted one of his machines while the other
one had a shared vg varied on, which created ghost disks on the
booted machine. Then when the customer tried to increase the
size of an lv through C-SPOC, the node with the ghost disks
could not find the related vg because of the ghost disks and
so the function failed on that node.

PROBLEM SUMMARY:
The customer had booted one of his machines while the other
one had a shared vg varied on, which created ghost disks on the
booted machine. Then when the customer tried to increase the
size of an lv through C-SPOC, the node with the ghost disks
could not find the related vg because of the ghost disks and
so the function failed on that node.

PROBLEM CONCLUSION:
An appropriate test for the existence of ghost disks and
a call to cl_disk_available to clean them up was added into
both clupdatevg and climportvg and the set -u instruction
was removed from cl_disk_available.

------

APAR: IY16857 COMPID: 5765E6400 REL: 220
ABSTRACT: SCRIPT/LINKS FROM GEO 2.1 CMD NAMES TO NEW GEO 2.2 CMD NAMES

PROBLEM DESCRIPTION:
HAGEO and GeoRM have had commands from 2.1 were superceded by
later versions with difference names in HAGEO and GeoRM 2.2 and
the old command names are no longer valid in 2.2.

PROBLEM CONCLUSION:
Scripts have been added such that the new commands can be
accessed via the old command name. If additional options are
added to the new commands in the future, one should use the
official 2.2 version command name to access the new option.

------

APAR: IY16960 COMPID: 5765D5100 REL: 311
ABSTRACT: BACKUP ID NOT UPDATED IN SDR WHEN BACKUP TIMES OUT DURING

PROBLEM DESCRIPTION:
This apar addresses the aftermath of a backup node not
responding during fence/unfence processing. The backup
node fails to ACK to a DEVICE_DB_UPDATES command;
the backup is fenced and a new backup is chosen. However,
the SDR is not updated to reflect the new backup id.

PROBLEM SUMMARY:
When a primary backup node times out during an Efence or
Eunfence operation, it is fenced off the switch and a new
backup node is chosen, but the SDR is not updated to
reflect the new backup's id.

PROBLEM CONCLUSION:
The error recovery Estart that gets invoked to handle
errors found during Efence or Eunfence will call the
function which will update the backup's name in the SDR.

------

APAR: IY17032 COMPID: 5765E5400 REL: 440
ABSTRACT: MOUNT POINT NOT REMOVED WHEN REMOVING FILESYSTEM WITH C-SPOC -

PROBLEM DESCRIPTION:
When the customer used C-SPOC to remove a filesystem and
specified to also remove the mount point, the mount point was
removed only on the node where the vg was varied on and not
on any of the other cluster nodes participating in the volume
group.

PROBLEM CONCLUSION:
If -r flag do try_parallel of rmdir mount point.

------

APAR: IY17164 COMPID: 5639I3900 REL: 310
ABSTRACT: DCE LIBRARY PROVIDING ERRONIOUS DEFINITION OF VSNPRINTF

PROBLEM DESCRIPTION:
DCE library providing erronious definition of vsnprintf.

LOCAL FIX:
Fixed by Transarc merge.

------

APAR: IY17216 COMPID: 5639I3700 REL: 310
ABSTRACT: RE-ENABLE CDSALIASING COMMANDS

PROBLEM DESCRIPTION:
A cdsalias delete will return with a 'not supported in
this release' error.

------

APAR: IY17217 COMPID: 5639I3700 REL: 310
ABSTRACT: PURIFY: FOUND STATIC OVERRUN IN CDSD INITIALIZATION

PROBLEM DESCRIPTION:
Purify: static overrun in CDSD initialization

------

APAR: IY17224 COMPID: 5639I3700 REL: 310
ABSTRACT: USER LEVEL CLIENT RPC CONTACTING KRPC CAUSES A CANCEL PACKET

PROBLEM DESCRIPTION:
AFS adapt server shows a communication timeout
using the AFS 'dd' command. Executing an
iptrace shows a RPC cancel packet sent to
the server (the server is a DFS in this
particular case).

------

APAR: IY17230 COMPID: 5639I3700 REL: 310
ABSTRACT: DCE_LOGIN CORES IN SOLARIS WHEN DEBUG_SEC IS SET OTHER THAN 1

PROBLEM DESCRIPTION:
JP Morgan had a core dump when DEBUG_SEC is set and did
dce_login. It worked only for DEBUG_SEC=1 and cored for
all other values (2-9). It works fine in AIX.

LOCAL FIX:
The problem is in binding_cache_lookup in
internal_binding.c. In the routine when trying
to print a pointer which is null Solaris cores.

------

APAR: IY17232 COMPID: 5639I3500 REL: 310
ABSTRACT: DCECP XATTR COMMANDS BROKEN FOR HOST SCHEMA

PROBLEM DESCRIPTION:
The "dcecp xattrschema" commands are broken for the host schema
objects (i.e. /.:/hosts/<hostname>/config/xattrschema).
dcecp> xattr catalog \
   /.:/hosts/sollasollew.austin.ibm.com/config/xattrschema
Error: msgID=0x17122157 invalid attribute name specified
dcecp> xattr create /.:/$_h/config/xattrschema/food -enc \
> printstring -aclmgr {srvrconf r r r r}
msgID=0x17122152 invalid acl manager type

------

APAR: IY17243 COMPID: 5765E5400 REL: 440
ABSTRACT: CLSMUXPD FAILS TO OPERATE PROPERLY IF NOFILES IS GREATER THAN

PROBLEM DESCRIPTION:
When nofiles is set to a value greater than 2000,
clsmuxpd does not operate properly.

LOCAL FIX:
set nofiles in /etc/security/limits to 2000 or less.

PROBLEM CONCLUSION:
modify clsmuxpd so that it will correctly handle file
descriptors.

------

APAR: IY17300 COMPID: 5639I3700 REL: 310
ABSTRACT: CLEARINGHOUSE CREATE HANGS AFTER DISABLE

PROBLEM DESCRIPTION:
dcecp -c clearinghouse disable
dcecp -c clearinghouse create with the same
name will hang.

PROBLEM CONCLUSION:
It seems the cdsd code is backgrounding some tasks and
getting stuck by Transarc merge. Removed the backgrounding
code.

------

APAR: IY17307 COMPID: 5639I3700 REL: 310
ABSTRACT: KILL OF KINIT CAN LEAVE 0-LENGTH CRED FILE

PROBLEM DESCRIPTION:
Customer was experiencing problems with DFS due to
zero-length credential files being created. These
files were created when an aixterm running a script
which invoked the kinit was closed. The kinit did
not recognize that no input was provided, and
proceeded as if a zero length password had been
entered. This resulted in a re-init of the cred
file (zeroing it). Usually, the kinit proceeds to
put the identity info in the cred file, followed by
any tickets it was able to refresh. In this case,
the kinit was killed by the exiting aixterm before
even the identity info was put into the cred file.

PROBLEM SUMMARY:
Zero length credential files can be created when kinit is
terminated. In particular, this occurred when kinit was
being called from a shell, and the shell was terminated.
When the shell terminated, stdin for kinit was closed, and
kinit behaved as if zero-length passwd was entered. It
re-inited the credentials file (erasing it), but before it
added the identity info to the cred file, it was terminated
as the calling shell exited.

PROBLEM CONCLUSION:
Putting in a check that will verify that something has been
entered as input. The input char is initialized to '0',
and if the read returns and the char is still '0', an
EINVAL is returned, since this was essentially a read
failure for kinit, and should not be treated as if the user
had input a newline ('n') indicating a zero-length passwd.

------

APAR: IY17308 COMPID: 5639I3700 REL: 310
ABSTRACT: CDS FIXES FROM TRANSARC

PROBLEM DESCRIPTION:
This defect is to fix the following problems from Transarc
merge
claire-24896-cdsd-inc-fd-rlimit-to-hard-max-at-startup
claire-24083-cdsd-prevent-replica-update-dir-delete-
deadlock
claire-22902-cdsd-dont-add-duplicate-lines-to-cds_files
claire-23604-cdsclerk-dont-process-empty-tower-set-in-
rankcompute
sonali-22215-cdsd-small-memory-leaks-in-cdsd

PROBLEM CONCLUSION:
Merge the transarc fixes into the dce310 and dcedev trees.

------

APAR: IY17311 COMPID: 5639I3700 REL: 310
ABSTRACT: SECD TRAPS AFTER "CANNOT BIND SERVER SOCKET ON PORT 88".

PROBLEM DESCRIPTION:
When configuring DCE server on Windows 2000 with KDC
running. Secd got an error "Cannot bind server socket on
port 88", because the Microsoft KDC running
on Windows 2000 server. But after a while, SECD died. This
might not happen in AIX becauses AIX uses different
functions. But this might happen in Solaris.

PROBLEM CONCLUSION:
n_udp_ports & max_udp_ports should be reset to 0 after
free(udp_port_nums) in closedown_network().

------

APAR: IY17313 COMPID: 5639I3700 REL: 310
ABSTRACT: DCE SECURITY SERVER CORES IN 'DCE_FREE'

PROBLEM DESCRIPTION:
The DCE security server cores with the following
stack:
(dbx) where
dce_free(??) at 0xd021a26c
libdcelibc_r.free(??) at 0xd0338488
trl_ops.dce_aud_close(??, ??) at 0x100d6e98
start_listeners(??) at 0x10002dac
_pthread_body(??, ??) at 0xd0208f58
The problem is the security server incorrectly
contains 2 signal handler threads. Therefore,
when a signal is sent to the process one of
the 2 threads will be randomly woken up. In
this particular case, the one woken up incorrectly
cleans up the security audit trailer. This audit
trailer should only be cleaned up in
'start_listeners after the rpc runtime engine has
been destroyed.

------

APAR: IY17326 COMPID: 5639I3700 REL: 310
ABSTRACT: BUILD: 001107A/DCEDEV/POWER,SPARC SECURITY.SRC

PROBLEM DESCRIPTION:
When building libdce.a following is output found in
libdce/shr.o.error, a symbol which appears
in src/security/client/login/sec_login_pvt.c
ld: 0711-317 ERROR: Undefined symbol: .init_v1_1_info

PROBLEM CONCLUSION:
Fixed the file src/security/client/login/sec_login_pvt.c

------

APAR: IY17343 COMPID: 5639I3700 REL: 310
ABSTRACT: MERGE IN TRANSARC SERVICE FIXES

PROBLEM DESCRIPTION:
Merge Transarc defects in DCE 3.1 code base.
24945 - sec_login_util-change-fopen-to-open
24659 - sec-login-delete-all-the-nc-creds-files
24414 - purge-handle-bad-ticket-cache
22830 - newgroups-better-initialization
23824 - sec-login-pvt-check-bounds-before-bcopy
23985 - sec-dont-remove-ffffffff-creds-if-in-prev_caches
24179 - update-identity-exp-more-often

PROBLEM SUMMARY:
Transarc code merge.

PROBLEM CONCLUSION:
Merged Transarc defects to DCE 3.1 codebase

------

APAR: IY17345 COMPID: 5639I3700 REL: 310
ABSTRACT: MERGE TRANSARC SOLARIS RPC FIX

PROBLEM DESCRIPTION:
Customers might get RPC errors.

PROBLEM CONCLUSION:
Merge TRANSARC solaris rpc fix.

------

APAR: IY17347 COMPID: 5639I3700 REL: 310
ABSTRACT: RETURN CODE OF READDIR_R CAUSES PROBLEM

PROBLEM DESCRIPTION:
readdir_r return code causes rmxcred infinite loop in
AIX 5.1.1. readdir_r which is a libc_r routine returns 0
even when the end of dir entry is encountered in
AIX 5.1.1 release. For previous release (for example
AIX 4.3.3 or 4.3.2), readdir_r returns
9 when no more file entry can be read in the directory.
The rxmcred checks the none-zero return code for a while
loop. So it will end up with an infinite loop for AIX 5.1.1.

PROBLEM CONCLUSION:
An extra check for a NULL file entry pointer was be put in
the code.

------

APAR: IY17348 COMPID: 5639I3700 REL: 310
ABSTRACT: TRANSARC MERGE FOR DCECP.SRC

PROBLEM DESCRIPTION:
dcecp errors.

PROBLEM CONCLUSION:
Transarc merge for dcecp.src

------

APAR: IY17349 COMPID: 5639I3700 REL: 310
ABSTRACT: AUDIT TESTCASE "API_FILTER" ON WIN2000: 2 ASSERTIONS FAILED

PROBLEM DESCRIPTION:
The audit testcases failed because of a bug in the audit
code. When the "world" filter is deleted, the audit code
does not delete the filter from its cache. Subsequent audit
AP Is were returning incorrect information for the "world"
filter guides.

PROBLEM CONCLUSION:
The error is on lines 364 and 366 in esl_file_read.c
     364: world_guides = NULL;
     366: world_overridable_guides = NULL;
These two lines should be modified to correctly set the
filter guides to NU LL as follows:
     364: *word_guides = NULL;
     366: *word_overridable_guides = NULL;

------

APAR: IY17350 COMPID: 5639I3700 REL: 310
ABSTRACT: RETURN FLAGS IN GSS_INIT_SEC_CONTEXT() DO NOT REFLECT REALITY

PROBLEM DESCRIPTION:
Client failed to get a forwardable ticket when Kerberos
delegation is being used.

PROBLEM CONCLUSION:
Return flags in gss_init_sec_context() do not reflect
reality.

------

APAR: IY17351 COMPID: 5639I3700 REL: 310
ABSTRACT: AUDIT API_LOG TESTCASE FAILS ON WIN 2000.

PROBLEM DESCRIPTION:
The testcases fail with error codes returned from
"dce_aud_prev" and "dce_aud_last" APIs.

PROBLEM CONCLUSION:
There was a bug in the "dce_aud_prev" API (evt_read.c)
which caused this problem. Memory was freed and being
accessed which resulted in undefined values assigned
to variables. The pointer "ifile" was set to "tfile"
(ifile = tfile). The pointer "tfile" was freed and then
"ifile" was accessed. This resulted in attempting to read
a large number of bytes (in millions) from a trail file
which only contained a few hun dred bytes.
Code changes will be made to free all the memory allocated
in the "while" loop at the end of all processing.

------

APAR: IY17376 COMPID: 5639I3500 REL: 310
ABSTRACT: MEMORY LEAK IN SECD

PROBLEM DESCRIPTION:
Liberty Mutual was seeing growth in sec
 After instrumenting it with debugmem and analyzing
found that defect 51301 introduced memory leak.

LOCAL FIX:
Use rps_ss_allocate instead of malloc in lm_copy_journal.

PROBLEM SUMMARY:
Liberty Mutual was seeing growth in secd. After
instrumenting it with debugmem and analyzing, found out that
fix for defect 51301 introduced memory leak.
Also in rpc__cn_assoc_send_frag if the call is cancelled,
memory is not freed.

PROBLEM CONCLUSION:
In lm_copy_journal use rpc_ss_allocate instead of malloc.
This memory will be freed by the stub code.
In rpc__cn_assoc_send_frag free the memory if the call
is cancelled.

------

APAR: IY17379 COMPID: 5639I3700 REL: 310
ABSTRACT: RMDCEWEB -V OPTION DOES NOT WORK AS DOCUMENTED!

PROBLEM DESCRIPTION:
When rmdceweb uses the -v option, it is supposed to prompt
for all the paramemters that weren't passed on the command
line.

PROBLEM CONCLUSION:
It was fixed by setting the WHICHPART environment variable
in rmdceweb to nothing instead of to 'all'.

------

APAR: IY17384 COMPID: 5639I3700 REL: 310
ABSTRACT: TRANSARC DEFECTS MERGE

PROBLEM DESCRIPTION:
-DCED ran out of file descriptors in the libpthread code.
-DCED crash in in rpc_ss_mem_free.

PROBLEM CONCLUSION:
Merge transarc defects 20320 into DCE 3.1 code base.
24925 - Increase the soft fd limit.
24825 - Fix several errors in hostdata_get().
20480 - Fixed the logic to detect gaps in sequence number.

------

APAR: IY17385 COMPID: 5639I3700 REL: 310
ABSTRACT: MERGE TRANSARC SERVICE FIXES

PROBLEM DESCRIPTION:
- secd core on failed fopen
- Corruption of flattened epac data caused secd crash

PROBLEM CONCLUSION:
Fixed the overwrite problem.

------

APAR: IY17386 COMPID: 5639I3700 REL: 310
ABSTRACT: NEED TO ADD CHANGED FILESET NAMES

PROBLEM DESCRIPTION:
When DirectTalk for AIX installs, it pre-reqs DCE filesets.
DCE 3.1 changed the fileset names and so install fails.

PROBLEM CONCLUSION:
Added dummy filesets, so that DirectTalk can install.

------

APAR: IY17391 COMPID: 5765E5400 REL: 440
ABSTRACT: ADDING/CONFIGURING AEN FAILS WITH SDD/DPO INSTALLED

PROBLEM DESCRIPTION:
Configuring (adding) Automatic Error Notification fails with
the following messages echoed to the SMIT screen: sili:
Operation failed. tech: Operation failed. dsh: 5025-509 tech
rsh had exit code 1

PROBLEM CONCLUSION:
The behavior of AEN with SDD
has been documented in the relevant chapters
of the HACMP Administration Guide and the
HACMP/ES Guide. Information was reviewed by
Venky. Information will also be added to
the Release Notes for version 4.4.1.

------

APAR: IY17403 COMPID: 5765E6400 REL: 220
ABSTRACT: GEO_VERIFY CORE DUMP

PROBLEM DESCRIPTION:
geo_verify core dumps from smit and the command line depending
on what options are supplied.

PROBLEM CONCLUSION:
Change geo_verify messages from the gmd msg catalog to the
gv msg catalog

------

APAR: IY17436 COMPID: 5639I3700 REL: 310
ABSTRACT: CAN'T DEBUG FUNCTIONS IN SRC/ADMIN/DCECP/TCPINFO.C

PROBLEM DESCRIPTION:
When debugging the core of registry migrate code, could not
look at any of the variables because dbx was getting
confused.

PROBLEM CONCLUSION:
Add dcp_ in front of all the functions in tcpinfo.c.

------

APAR: IY17458 COMPID: 5639I3500 REL: 310
ABSTRACT: DCECRED_XXXXXXXX.DATA.DB FILE CURRUPTED, DFSBIND DUMPS CORE

PROBLEM DESCRIPTION:
IBM Rochester is seeing a problem where dfsbind core
dumps due to a corrupted dcecred_XXXXXXXX.data.db file.

LOCAL FIX:
To open and lock the .data.db file, we call dce_db_open_file(),
passing in the database filename and flags indicating how the
file should be locked. dce_db_open_file() calls dbopen() to
open the database and then dce__db_lock_file() to lock it. That
looks fine on first glance. But looking more closely at
dbopen(), we see that not only does it open the file, it also
reads data out of the file. This is a Bad Thing. The file
should be locked immediately after the open(), and before any
reads or writes.
So, the fix is to move the lock from dce_db_open_file() down
into the lower level open routines (__bt_open(), __hash_open()
and __rec_open()), immediately after they do the open(). To do
this, dbopen() and __{bt,rec,hash}_open() had to be modified to
pass in the flags argument that had the lock flags. And a
slightly different interface to dce__db_lock_file() had to be
written that takes the file descriptor instead of the database
handle as an argument (the database handle is not constructed
until after the call to dbopen()).

PROBLEM SUMMARY:
One of the credential files (dcecred_NNNNNNNN.data.db)
associated with a users login context can become corrupted.
This can cause any DCE process that tries to access that
file to dump core. These may include klist, kinit,
cdsclerk, dfsbind and others.

PROBLEM CONCLUSION:
Fixed the file locking logic so that the file is always
locked immediately after it is opened, and before it is
read.

------

APAR: IY17581 COMPID: 5765D5100 REL: 311
ABSTRACT: FSD RESPONSE TIME

PROBLEM DESCRIPTION:
fsd response time

PROBLEM SUMMARY:
When a node is heavily loaded, the switch daemon can be
delayed processing certain packets. In some cases, the node
is dropped from the switch.

PROBLEM CONCLUSION:
The packet processing code in the fault service daemon has
been changed to improve processing time.

------

APAR: IY17684 COMPID: 5765E5400 REL: 440
ABSTRACT: HACMP 4.4 HAS AND CRM SUPPORT FOR AIX5.1

PROBLEM DESCRIPTION:
HACMP 4.4 Support of AIX Version 5.1
This APAR adds support for AIX Version 5.1 to the HAS
and CRM features of HACMP 4.4.
Without these changes cluster synchronization of topology
and the use of the lock manager will fail.

PROBLEM SUMMARY:
HACMP 4.4 Support of AIX Version 5.1 for HAS and CRM.
This APAR includes required updates to support AIX Version .
Without these changes cluster synchronization of topology
and the use of the lock manager will fail.

PROBLEM CONCLUSION:
HACMP Support for AIX Version 5.1
This APAR adds support of AIX Version 5.1 to the HAS
and CRM features of HACMP 4.4.
To allow cluster synchronization on this level of AIX, the
code was modified so that when PVIDs or VGIDs are retrieved,
they are padded out to 32 bytes if they are not already and
to always use 32-byte comparisons for PVIDs and VGIDs.
For lock manager support on the new AIX release the code
was updated to replace the call to msgrcv with kmsgrcv.

------

APAR: IY17686 COMPID: 5765E5400 REL: 440
ABSTRACT: UNABLE TO SYNC CLUSTER TOPOLOGY DUE TO CLLOG ERROR MESSAGE

PROBLEM DESCRIPTION:
During verification of log files, cllog will fail, issuing the
erroneous message that cluster.log has already been redirected.

PROBLEM SUMMARY:
During verification of log files, cllog will fail, issuing the
erroneous message that cluster.log has already been redirected.
The real cause of the problem is an incorrect number of ''
characters in the following awk calls:
awk '/local0.info/ { print $2 }' /etc/syslog.conf
awk '/user.notice/ { print $2 }' /etc/syslog.conf
There should be three '' characters before the '$', not one.

PROBLEM CONCLUSION:
The code was changed to pass grep command which does not
require any escaped characters to cl_rsh so that the entire
line is returned to the local machine. Once returned the
line is then echoed into awk to print just the $2 field.

------

APAR: IY17996 COMPID: 5765E6400 REL: 220
ABSTRACT: HAGEO: GEO_VERIFY NEEDS TO CHECK LV SIZE FOR >32GB

PROBLEM DESCRIPTION:
If a lv of size greater than 32G is used it will not be geo-
mirrored but there is no check on this in the current system

PROBLEM CONCLUSION:
add code to geov_chks.c code to check if the raw device exceeds
the limit

------

APAR: IY17997 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM: MKDEV -L <GMD> RETURNS CONFUSING, INACCURATE ERROR MSG

PROBLEM DESCRIPTION:
mkdev -l gmd returns a non-useful message.

PROBLEM CONCLUSION:
Change error code returned from startgmd such that it can be
better understood by mkdev.

------

APAR: IY17998 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM: TWO CONFLICTING SETS OF COMMAND LINE UTILS

PROBLEM DESCRIPTION:
HAGEO and GeoRM have had commands supercededby later vers with
difference names. The old commands should point to the newer
versions of these commands.

PROBLEM CONCLUSION:
The old scripts add_geonode and add_geosite have been replaced
with scripts which call the newer versions of these programs
with the correct command line arguments.

------

APAR: IY17999 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM: GMDSIZING W/BAD COMMAND LINE OPTS - CRYPTIC ERRORS

PROBLEM DESCRIPTION:
gmdsizing prints out a message when no parameters are supplied.

PROBLEM CONCLUSION:
Make gmdsizing print out the usage message if no command line
arguments are supplied.

------

APAR: IY18000 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM:GMDSTAT EXITS 0 EVEN AFTER FAILURE

PROBLEM DESCRIPTION:
gmdsizing should exit with a return code other than
zero when incomplete or invalid parameters are given.

PROBLEM CONCLUSION:
return a code other than zero in the event of imcomplete or
inaccurate parameters being supplied.

------

APAR: IY18001 COMPID: 5765E6400 REL: 220
ABSTRACT: HAGEO: CANNOT CHANGE A GMD TYPE VIA SMIT

PROBLEM DESCRIPTION:
When a remote host is removed from the configuration using
smit there can be a problem with HAGEO determining that the
remote string sent means that the host should be removed. It
instead tries to parse that string and cannot find a host name.

PROBLEM CONCLUSION:
The chggmd command should ignore remote hosts which have zero
length. This means that they are from smit and should be
disgarded as the user has most likely removed the definition.

------

APAR: IY18002 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM: SMIT CREATE NODE ACCEPTS INVALID PROMOTE FAILURE TIMEOUT

PROBLEM DESCRIPTION:
geoRM should not accept invalid values for the promote
failure timeout value.

PROBLEM CONCLUSION:
Add a check for the size of the promote failure timeout
paramter. We have set limits of between 0 and 1800sec
(30mins).

------

APAR: IY18003 COMPID: 5765E6400 REL: 220
ABSTRACT: SMITTY CHGMD CAUSES DEFINITION ON 2ND REMOTE PEER TO BE LOST

PROBLEM DESCRIPTION:
When attempting to change a characteristic of a GMD for which
two remote peers are defined, the command executed by the smit
menu deletes the second remote peer from the GMD definition.
The GMD definition looks correct on the Change/Show GMD smit
screen (ie 2 remote peers are listed on the screen)
but when the command is executed, smit.script shows that the
second remote peer is cut off from the executed command.
The result is that only one remote peer is defined after using
the Change GMD command

PROBLEM CONCLUSION:
SMIT menus need adjusting with the required field being
set to '?' as opposed to being left blank.

------

APAR: IY18004 COMPID: 5765E6400 REL: 220
ABSTRACT: LOGIC ERROR IN CHANGE SITE - SITE DELETED FROM GEONODE

PROBLEM DESCRIPTION:
If a (valid) change is made to the site membership and the
topology resynchronized, HAGEO Verification incorrectly
reports errors.
In addition, the error messages it returns do not make sense:
ERROR: orion - Node has site (), but site siteA
is configured with the node.
ERROR: Node sigmund, Node orion is configured at
site , (locally node is defined at siteA)

PROBLEM CONCLUSION:
When updating the GEOnode class, geo_change_site should look to
see if that node required updating. If it doesn't then no
changes should be made in the GEOnode class for that node.

------

APAR: IY18005 COMPID: 5765E6400 REL: 220
ABSTRACT: HAGEO: SELECT NODES FROM LIST-SMIT REPORTS A TYPO

PROBLEM DESCRIPTION:
In the Change/Show Site screen, if you use the pick list to
select Site Nodes, smit will return a warning that a node
does not exist, but the node name it warns about is not
a name that you selected.
It appears that the selected node names are written over the
previous values. If the new nodes have shorter names than
the previous ones, the part of the field that was not
overwritten stays in the field (even though it is not
visible to the user)
This is probably a general SMIT problem that exists in other
screens as well.

PROBLEM CONCLUSION:
There is a non-terminated string. This causes the old node
definition to be left in the newly parsed string.

------

APAR: IY18006 COMPID: 5765E6400 REL: 220
ABSTRACT: HAGEO 2.2 AND GEORM 2.2 CANNOT REPLACE LINKS IN A PTF

PROBLEM DESCRIPTION:
Hard links cannot be replaced in a PTF release.

PROBLEM CONCLUSION:
Hard links will be left in tact.

------

APAR: IY18008 COMPID: 5765E6400 REL: 220
ABSTRACT: HAGEO: CAN NOT CHANGE PROMOTE FAILURE TIMEOUT VIA SMIT PANEL

PROBLEM DESCRIPTION:
The commands to change the promote failure timeout
do not work correctly nor do the associated smit panels.

PROBLEM CONCLUSION:
The smit panels have been corrected and the command
geo_change_node updated to handle the promote failure timeout
correctly.

------

APAR: IY18018 COMPID: 5765E6400 REL: 220
ABSTRACT: MULTIPLE PROBLEMS WITH HAGEO AND GEORM 2.2

PROBLEM DESCRIPTION:
HAGEO and geoRM have multiple problems in commands.
geo_show_config, does not always display the site information,
which can cause a problem when using smit.

PROBLEM CONCLUSION:
Rectify problems in geo_show_config, geo_add_node and
geo_change_node.

------

APAR: IY18019 COMPID: 5765E6400 REL: 220
ABSTRACT: GEO_MOUNT_FS FAILS - INCORRECT USAGE

PROBLEM DESCRIPTION:
In the efix of HAGEO sent March 9, Geo_mount_fs fails and
gives a usage statement in hacmp.out.
Filesystems for GMD's are not mounted.
This problem does not occur in HAGEO 2.2 gold + PTF 2

PROBLEM CONCLUSION:
The show_geonode and show_geosite commands have been replaced
by geo_info and geo_show_config commands.

------

APAR: IY18022 COMPID: 5765E6400 REL: 220
ABSTRACT: GEORM: FAILING ONE NETWORK OF TWO STOPS DATA REPLICATION

PROBLEM DESCRIPTION:
If one of two networks fails then HAGEO should continue to
mirror data over the other network. Once this network is
restored then data should be mirrored over both networks.
If the second network is then failed, there can be a problem
with the load balancing algorithm getting into a condition
that prevents the load from being re-distributed onto the
surviving network.

PROBLEM CONCLUSION:
The algorithm which balances the load over two or more networks
has been modified such that the priority of a network which
is not being used it changed constantly. The change is only
small, but this ensures that it will eventually be tried again.

------

APAR: IY18033 COMPID: 5765E5400 REL: 440
ABSTRACT: CLDARE -T FAILS ON CLUSTER WITH ENHANCED SECURITY

PROBLEM DESCRIPTION:
cldare -t fails with error "godm: Failed to get tickets"
on cluster with enhanced security and if no principal for host.

PROBLEM SUMMARY:
After cl_setup_kerberos completes without error and the cluster
security is set to enhanced cluster sync will fail with
following errors,
get_tickets: krbtgt kerberos authentication failed.
2504-008 Kerberos principal unknown
godm: Failed to get tickets

PROBLEM CONCLUSION:
The get_tickets function should not use host name for
principal. Change gethostname to use cluster interface name.

------

APAR: IY18036 COMPID: 5765E5400 REL: 440
ABSTRACT: ERROR APPLICATION SERVER PARAMETERS CANNOT CONTAIN SPACES OCCURS

PROBLEM DESCRIPTION:
Cluster Planing Worksheets Application Server has entries for
scripts. If parameters follow the script name an error message
Application Server parameters cannot contain spaces is sent.

PROBLEM CONCLUSION:
Change error checking to allow spaces.

------

APAR: IY18038 COMPID: 5765E5400 REL: 440
ABSTRACT: HACMPDISKTYPE ODM CLASS NOT ACCESSIBLE FROM EVENT SCRIPTS

PROBLEM DESCRIPTION:
Typically, spurious error messages about the class not being
found. In the cases where disk customization is used, it is
ineffective.

PROBLEM CONCLUSION:
Add HACMPdisktype to the list of ODM classes moved from DCD to
ACD

------

APAR: IY18039 COMPID: 5765E5400 REL: 440
ABSTRACT: CSPOC ADD USER VERY SLOW

PROBLEM DESCRIPTION:
cspoc add user very slow because if many users in /etc/passwd
(>7000) the "lsuser -a id ALL" will use 2 minutes of CPU time
on a RS/6000 model F50.

PROBLEM SUMMARY:
If system has >5000 users CSPOC mkuser can take two minutes.

PROBLEM CONCLUSION:
Replase lsuser with awk on /etc/passwd.

------

APAR: IY18040 COMPID: 5765E5400 REL: 440
ABSTRACT: CLFINDRES DOES NOT WORK ON INACTIVE NODES

PROBLEM DESCRIPTION:
cannot use clfindres on an inactive node.

PROBLEM SUMMARY:
Cannot us clfindres on an inactive node

PROBLEM CONCLUSION:
Forward resource group location request to active node.

------

APAR: IY18043 COMPID: 5765E5400 REL: 440
ABSTRACT: THE NFSO -H DUMP_DUPCACHE TAKES ONLY ONE LOGDEV

PROBLEM DESCRIPTION:
When multiple log devices are passed an informational
message is generated.

PROBLEM CONCLUSION:
Change to only use first log device on volume group.

------

APAR: IY18044 COMPID: 5765E5400 REL: 440
ABSTRACT: AFTER MKSYSB RESTORE, UNAME DOES NOT MATCH HOSTNAME

PROBLEM DESCRIPTION:
After mksysb restore uname does not match hostname.

LOCAL FIX:
Execute 'uname -s $(hostname)'.

PROBLEM SUMMARY:
Cluster must be brought up twice to have cluster operate
properly.

PROBLEM CONCLUSION:
It appears mksysb has changed. Rather than figuring out what
to do to fix mksysb. A minor change was added to harc.net.sh.

------

APAR: IY18045 COMPID: 5765E5400 REL: 440
ABSTRACT: HAS: OLPS APPLICATION SHOULD PREVENT CWOF RG WITH >2 NODES

PROBLEM DESCRIPTION:
The online planning sheet application allows the user to
define a cascading without fallback resrouce group which
contains more than two nodes. However, this is not permitted
in classic HAS.

PROBLEM CONCLUSION:
Added code to the method
PanelResourceGroups.resetParticipatingNodes such that, if the
number of nodes defined to a group is larger than two and the
kind of cluster is HACMP, the CWOF parameter is set to false
and the CWOF radio button panel is rendered invisible.

------

APAR: IY18047 COMPID: 5765E5400 REL: 440
ABSTRACT: CL_CCVG TASKGUIDES SHOULD WORK WITH C LOCALE

PROBLEM DESCRIPTION:
If NLS support is being used and LC_ALL is set to point to
non-english message catalogs, output to the log files will
also be in the non-english language. Log files should be in
english to allow for more effect customer support.

PROBLEM CONCLUSION:
A previous fix attempted to set LC_ALL=C, but used improper
Perl script syntax. Add the missing semi-colon.

------

APAR: IY18048 COMPID: 5765E5400 REL: 440
ABSTRACT: HACMP.OUT AND CSPOC.LOG HAS TRANSLATED MESSAGE

PROBLEM DESCRIPTION:
If LANG is set to point to a non-english message catalog, the
HA log files may contain non-english text. This can make it
difficult to receive customer support in the event of problems.

PROBLEM CONCLUSION:
This is a change from the solution posted above:
Add 'LANG=C' on the command line when clgetif is called.
De-internationalize clsetenvgrp. This program is not intended
to be run by users, and thus none of its output should be
showing up anywhere other than the logs. The command does not
have a man page or any other user documentation, which means
that this solution should be appropriate.

------

APAR: IY18054 COMPID: 5765E5400 REL: 440
ABSTRACT: HACMP: MISC TYPOS

PROBLEM DESCRIPTION:
Upon finding a typographical syntax error in a Korn Shell
script, all ksh scripts under /usr/sbin/cluster were executed
with "ksh -n" to check for syntax errors. Additionally they
were checked for trailing backslashes that are followed by a
space or a tab.

PROBLEM CONCLUSION:
Added missing quotes

------

APAR: IY18055 COMPID: 5765E5400 REL: 440
ABSTRACT: IF CLUSTER NODE FORCED DOWN THEN CLSTART THE LOCATION OF THE

PROBLEM DESCRIPTION:
In a two node cluster with one rotating resource group if a the
node forced down has the resource group then HA is started
again on that node the location of the resource group is not
known by the cluster.

PROBLEM SUMMARY:
Rotating resource groups can be brought online on more than one
node through node migration (cldare -M). If the node holding
the rotating resource group is forced down, "clfindres" shows
that RG to be down, even when cluster services are restared on
that node. If the RG is subsequently migrated to another node,
no attempt will be made to stop the resources on the owning
node and both nodes will hold the resources.

PROBLEM CONCLUSION:
clsetenvgrp needs to be changed to output rotating
resource groups when the cluster services on a node come up
after a forced down (the shared label belonging to the
rotating resource group is already UP). This will enable
an attempt to reacquire the resource groups.

------

APAR: IY18056 COMPID: 5765E5400 REL: 440
ABSTRACT: HAS:CL_SWAP_ATM_IP_ADDRESS NOT SHOWING CORRECTLY IN NETSTAT

PROBLEM DESCRIPTION:
HAS:cl_swap_ATM_IP_address not showing correctly in
netstat. On ATM Classical IP network, swap_adapter
events will run to completion with no error
indication as recorded by HACMP logs, however netstat
shows the address on the same interface after the
swap.
Results of PING test after the "swap" are as expected:
A PING of the service label is successful after the swap.
A PING of the standby label is not successful.
will not actually occur as seen using netstat.
Addtional netstat and hacmp.log information on request.

PROBLEM CONCLUSION:
the actual symptom, that an adapter swap is
not verifyable using netstat -i (as in the other
case of network types) is not changed.
Adapter swaps for networks of type ATM will keep
beeing different than for other network types, in
that the association network interface -IP address
will remain, however, the association ATM device
- network interface will change, hence in case of
a failure of the adpater the adapter swap will
will only be verifiable using the arp
command. However, during an adapter swap the
hardware address did not get moved, this is fixed
in this defect.
As well an incorrect cl_hats_adapter call is
removed.

------

APAR: IY18057 COMPID: 5765E5400 REL: 440
ABSTRACT: INCORRECT LIST OF SHARED FILESYSTEMS UNDER SMIT CHANGE / SHOW

PROBLEM DESCRIPTION:
When a Resource Group only contains one Node name in node list
and no
    filesystems nor Volume Groups are specified for this
    Resource Group, the smit menu: Change / Show
Characteristics of a Shared File System under the cspoc
options, displays an incorrect list of filesystems, including
all filesystems in rootvg.

PROBLEM SUMMARY:
When a Resource Group only contains one Node name in node list
and no filesystems nor Volume Groups are specified for this
Resource Group, the smit menu: Change / Show
Characteristics of a Shared File System under the cspoc
options, displays an incorrect list of filesystems, including
all filesystems in rootvg.

PROBLEM CONCLUSION:
Modify HACMP smit cspoc odm sm_cmd_opt and sm_cmd_hdr entries
to use correct flags.

------

APAR: IY18058 COMPID: 5765E5400 REL: 440
ABSTRACT: CLVER DISPLAYS ERROR WITH 3.2.0.6 SSP.BASIC: ERROR: SERVICE

PROBLEM DESCRIPTION:
clver incorrectly displays the following error when the
cluster includes an HPS network and another tcpip network
such as ethernet:
Service adapter <adapter name> is improperly configured on
node <node name>.
The problem appears to be caused by a naming convention
change in PSSP regarding the css adapter/interface.

PROBLEM SUMMARY:
cluster verification fails with following error:
ERROR: Service adapter <adapter name> is improperly configured
on node <node name>.

PROBLEM CONCLUSION:
Modify clver so that it is not so picky about the css name
it is looking for in the CuAt and CuDv ODM classes.

------

APAR: IY18734 COMPID: 5765D5100 REL: 311
ABSTRACT: LATEST PSSP 3.1.1 FIXES AS OF MAY 2001

PROBLEM DESCRIPTION:
This is the lastest PSSP ptf as of May 2001.
Order this apar to get all of the ptfs as of May 2001.

PROBLEM SUMMARY:
This is a packaging apar for PSSP 3.1.1 fixes
as of May 2001.

PROBLEM CONCLUSION:
This is a packaging apar for PSSP 3.1.1
fixes as of May 2001.

------

APAR: IY19015 COMPID: 5639I3700 REL: 310
ABSTRACT: DCE310 MAINTENANCE PACKAGE FOR PTF SET 4

PROBLEM DESCRIPTION:
THIS APAR CREATED FOR MAINTENANCE PACKAGE OF PTF SET 4
dce310 Maintenance Package for PTF SET 4

PROBLEM SUMMARY:
AIXDCE310 MAINTENANCE PACKAGE FOR PTF SET 4

PROBLEM CONCLUSION:
AIXDCE310 MAINTENANCE PACKAGE FOR PTF SET 4

------

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]