ptx®/EES, the Error/Event Subsystem (EES), improves the fault management, serviceability, and usability of NUMA-Q® systems by providing a single set of interfaces to report, log, and retrieve system errors and events.
These Release Notes contain the following information:
Compatibility information
Installation instructions
Problems fixed since the previous release
Open problem reports
ptx/EES is compatible with the following software products running on Symmetry® and NUMA-QTM systems:
DYNIX/ptx® V4.4.4 or greater
ptx/LAN V4.5.3 or greater
ptx/TCP/IP V4.5.1 or greater
ptx/BaseComms V1.1.0 or greater
ptx/EES is installed using ptx/ADMIN® and ptx/INSTALL.
ATTENTION We recommend that you install /var/ees on its own filesystem, otherwise the root filesystem may fill up quickly due to the size of the ptx/EES logfile. If you choose to make a new filesystem for /var/ees, we recommend that you create the new filesystem prior to installing the ptx/EES software.
The ptx/EES User's Guide contains more information about managing the size of the ptx/EES logfile.
Refer to the DYNIX/ptx and Layered Products Software Installation Release Notes for installation procedures.
239727. The kernel no longer panics when itx calls cmn_err.
237983. ees_notify no longer incorrectly removes an item from the action registry.
239860. ees_view is now installed with 6555 permission.
238459. ees_notify no longer incorrectly removes items from the action registry.
238352. Binary records are now correctly displayed.
238351. ees_notifyd now correctly stores logrec and binrec events.
238350. ees_notify no longer gives numerous false warnings regarding action scripts.
238169. ees_view -log -collate option now works.
240625. dl_ees_syslog function now works correctly.
240614. syslog events are now copied to the EES log.
240613. The severity mapping of syslog calls is now correctly handled.
240318. ees_view and ees_logmgr no longer hang when processing a corrupt log file.
239615. The ees_send API no longer generates multiple log records for the same error.
238476. ees_view now scans the EES log from the newest to the oldest records.
242328. Panic caused by %b format specifier in description string no longer occurs.
242510. Panic with page fault in eesdr_buffer_rec no longer occurs.
The number that appears in parentheses in problem report titles is the Problem Tracking System number assigned to the report. These problems will be corrected in a future release unless otherwise noted.
The eeslog is a binary log file with header information, unlike the other log files, which are ASCII text. If you use the Log File Management menu of ptx/ADMIN to clear the log file, the eeslog is backed up and the size is set to zero, but it does not retain the header information. This causes ees_view to display an error message.
Workaround. Do not use the Log File Management menu of ptx/ADMIN to clear out the eeslog file or for any other ptx/EES task. If you want to clear out the log file:
Stop the ees_notifyd and the eesd daemons by entering the following: sh /etc/init.d/ees stop
Manually cp the eeslog to a backup filename.
Restart the two daemons by entering the following:
sh /etc/init.d/ees start
Filters that use ees_time as an attribute tag cannot be properly registered because the : separators used to separate day, month, year and so on are confused with the : separators used in the filter registry. Once such a filter is added to the registry, it is never found by the utility. These filters do not, however, corrupt the registry, so other filters will work fine.
Workaround. Do not enter filters using the ees_time attribute. If one is entered, remove it manually.
The ees_get_logevent(3) man page has an incorrect summary description. The summary description should read "get an event from an EES log".
Workaround. None.
If you execute the command ees_notify_id -remove 2.5, any action with an ees_notify_id of 2 might be removed.
Workaround. Use only whole numbers for ees_notify_ids.
ees_view -out log.3 -lang log -log "log1.log2" -collate command gives an error message indicating that log.3 is an invalid log file.
Workaround. None.
The ees_close_prov(3) man page incorrectly lists ees_open_prov in the NAME field, when it should be ees_close_prov.
Workaround. None.
If you run ees_view -new in one window and then run it in another window and later kill it, a subsequent attempt to use ees_view will cause the original ees_view to fail with a memory fault.
Workaround. Do not use multiple instances of ees_view -new.
ees_logmgr does not properly lock for multiple instances and could possibly damage the ptx/EES log if the processes are trying to write to the log at the same time.
Workaround. Do not use multiple instances of ees_logmgr.
Some ptx/EES log entries can have a severity level of "none", yet this is not recognized as a valid value by the filter library.
Workaround. None.
The following two commands may return an non-zero exit code when called from the CommandPoint Admin product:ees_logmgr -compact and ees_logmgr -size. These non-zero codes may be returned even if the operation is successful.
Workaround. None.