Net Timers Logging Location Codes

CAUSE: A free-pool entry was needed for a timer-entry element, to hold information about a new net timer being obtained, but the pool was empty. A separate timer entry must exist for every net timer on the system.

ACTION: No new timer was obtained. The system is probably operating under a heavy load. Other processes on the system have used up all the free timer entries, or are starting timers with granularities so small as to use up all the pool entries for timer-intervals. Other processes on the system will probably start seeing this problem as well, though that should be a temporary condition. One possible workaround is to configure a higher value in NMMGR for the maximum number of TCP connections.

CAUSE: A free-pool entry was needed for an interval- head element, to link a net timer into a new interval, but the pool was empty. A new interval is required whenever there is no open interval positioned where a timer entry needs one.

ACTION: If the error occurred when getting a new timer, no new timer was obtained. If it occurred on a reset, the timer was not reset, and has been left stopped. The system is probably operating under a heavy load. Other processes on the system have used up all the free timer entries, or are starting timers with granularities so small as to use up all the pool entries for timer-intervals. Other processes on the system will probably start seeing this problem as well, though that should be a temporary condition. One possible workaround is to configure a higher value in NMMGR for the maximum number of TCP connections.

CAUSE: When releasing a system timer following a race-condition when that timer could not be cached, the system routine Release_timer reported an error. Timers must be released when the net timer cache is full or not configured (PARM = status from Release_timer).

ACTION: Unless there is corruption within system timers, this probably indicates a Net Timers bug. Another remote possibility is that the Net Timers port data has been corrupted from outside, so other, possibly serious problems may soon appear as well. In addition, one system timer may have been lost; if this continues, the system could eventually run out of system timers. Submit an SR.

CAUSE: While attempting to stop, restart, or reschedule an existing system timer, the system routine Reset_timer_in_heartbeats failed or reported a status which was not expected given the current state of Net Timers. Resetting a system timer is a normal operation done whenever the first interval in a timer list changes (PARM = status from Reset_timer_in_heartbeats, PORT = address of timer list).

ACTION: Unless there is corruption within system timers, this probably indicates a Net Timers bug. Another remote possibility is that the Net Timers port data has been corrupted from outside, so other, possibly serious problems may soon appear as well. All timers associated with the faulty system- timer are now stalled, and will probably never pop. Depending on your configuration, this may or may not affect all timers on that system's network. Submit an SR.

CAUSE: While attempting to obtain a new system timer, the system routine Get_timer_in_heartbeats failed. Getting a new system timer is done the first time each configured timer list is used, and also during normal processing of race conditions when no cached timer was available (PARM = status from Get_timer_in_heartbeats, PORT = address of timer list).

ACTION: Unless there is corruption within system timers, the system has probably run out of timers; this might be indicated by a PARM value of FFFE006E: no entry available. If the error occurred when getting a new timer, no new timer was obtained. If it occurred on a reset, the timer was not reset and has been left stopped; all timers to be associated with that system timer are now stalled, and will probably never pop. Depending on your configuration, this may or may not affect all timers on that system's network. Submit an SR.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to start a new net timer or restart an existing one, as part of a higher-level reset- or get-timer operation, it was determined that a new interval was needed, but the free pool was empty, preventing allocation of a new interval-head element. A new interval is required whenever there is no open interval positioned where a timer entry needs one. (PORT = net timer ID)

ACTION: Review information regarding the previous errors, for possible hints as to the cause. Other processes on the system have used up all the free entries in the pool.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to start a net timer as part of a higher-level reset- or get-timer operation, there was a problem resetting the system timer associated with that net timer and others. A timer list's system timer is normally reset whenever insertion of the new net timer changed the first interval in the list. (PORT = net timer ID).

ACTION: Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to reschedule a running net timer, the timer was successfully removed from the old interval it was running in, but there was a problem restarting it. To get the net timer running again, all steps necessary to attach it to a new interval must succeed. (PORT = net timer ID).

ACTION: Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: An attempt was made to free a net timer before or during Net Timers startup, or during or after shutdown. Timers cannot be freed unless Net Timers is completely up. (PORT = net timer ID).

ACTION: If this occurs during network shutdown, it can probably be ignored. If it occurs at any other time, submit an SR. Whether the net timer ID being freed was valid or not is not known.

CAUSE: An attempt was made to free a net timer having an ID of zero (PARM = 0).

ACTION: There is a bug in the calling module. Probably that module previously attempted to get a timer, but ignored an error reported by Net Timers and saved the returned ID of zero for later use. It is now attempting to free that nonexistent timer. This may lead to other errors or hangs in the calling module. Submit an SR.

CAUSE: An attempt was made to free a net timer for which a deferred release has already been posted. Deferred release is used only during the pop of a timer, since it cannot be freed until the trap routine exits, completing the pop. Since freed net timers can be immediately reallocated to anyone, they cannot be safely freed twice. (PORT = net timer ID).

ACTION: This probably means a Transport timer trap- routine has accidentally re-freed a timer it already freed. Another possibility is that a second processor has attempted to free the same timer while that trap routine was still executing; this might be unlikely due to timing constraints. If necessary, submit an SR.

CAUSE: An attempt was made to free a net timer which had already been freed, or a timer which was in some sort of bad state. Since freed net timers can be immediately reallocated to anyone, they cannot be safely freed twice. (PORT = net timer ID).

ACTION: If you suspect the timer has not been freed, another remote possibility is data corruption of the timer entry. If necessary, submit an SR.

CAUSE: An attempt was made to get a net timer before or during Net Timers startup, or during or after shutdown. Timers cannot be obtained unless Net Timers is completely up. (PORT = net timer ID).

ACTION: If this occurs during network shutdown, it can probably be ignored. If it occurs at any other time, submit an SR.

CAUSE: An attempt was made to set a new net timer to a negative time interval.

ACTION: No new timer was obtained. Negative intervals are not allowed. There is a bug in the calling module. Probably that module is doing its own timing by subtracting timestamps over a period of time, and has made an incorrect calculation. See "Submitting an SR" in appendix A of this manual.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to get a new net timer, it was discovered that the free pool was empty, preventing allocation of a new timer entry element. A separate timer entry must exist for every net timer on the system.

ACTION: Review information regarding the previous errors, for possible hints as to the cause. Other processes on the system have used up all the free entries in the pool.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to start a new running net timer, there was a problem getting it into a running timer list. For the new net timer to run, it must be successfully attached to an interval in a timer list having a running system timer.

ACTION: No new timer was obtained. Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: An attempt was made to reset a net timer before or during Net Timers startup, or during or after shutdown. Timers cannot be reset unless Net Timers is completely up. (PORT = net timer ID).

ACTION: If this occurs during network shutdown, it can probably be ignored. If it occurs at any other time, submit an SR. Whether the net timer ID being freed was valid or not is not known.

CAUSE: An attempt was made to reset a net timer having an ID of zero (PARM = 0).

ACTION: There is a bug in the calling module. Probably that module previously attempted to get a timer, but ignored an error reported by Net Timers and saved the returned ID of zero for later use. It is now attempting to reset that nonexistent timer. This may lead to other errors or hangs in the calling module. Submit an SR.

CAUSE: An attempt was made to reset an existing net timer to a negative time interval (PORT = net timer ID).

ACTION: The timer was not reset from its current condition, running or stopped. Negative intervals are not allowed. There is a bug in the calling module. Probably that module is doing its own timing by subtracting timestamps over a period of time, and has made an incorrect calculation. See Appendix A “Submitting an SR ” of this manual.

CAUSE: An attempt was made to reset a net timer for which a deferred release has been posted. Deferred release is used only during the pop of a timer, since it cannot be freed until the trap routine exits, completing the pop. Since freed net timers can be immediately reallocated to anyone, they cannot be safely reset after freeing. (PORT = net timer ID).

ACTION: This probably means a Transport timer trap- routine has accidentally reset a timer it already freed. Another possibility is that a second processor has attempted to reset the freed timer while that trap routine was still executing; this might be unlikely due to timing constraints. If necessary, submit an SR.

CAUSE: This is supplemental logging to a previous Net Timers error. While attempting to reschedule a running net timer, there was a problem removing the timer from the old interval it was running in, or a problem restarting it. Both steps must succeed for the rescheduling to work. (PORT = net timer ID).

ACTION: The timer has been left in an indeterminate state. Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: This is supplemental logging to a previous Net Timers error. A problem occurred while attempting to restart a stopped net timer. To run a previously- stopped net timer, the timer must be successfully attached to an interval in a timer list having a running system timer. (PORT = net timer ID).

ACTION: The timer has been left in an indeterminate state. Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: This is supplemental logging to a previous Net Timers error. A problem occurred while attempting to reschedule a popping net timer. By definition a popping net timer is already stopped, so to get it running again it needs to be successfully attached to an interval in a timer list having a running system timer. (PORT = net timer ID).

ACTION: The timer has been left in an indeterminate state. Probably a Transport timer trap routine was attempting to reset the timer. Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: An attempt was made to restart a net timer which had already been freed, or a timer which was in some sort of bad state. Since freed net timers can be immediately reallocated to anyone, they cannot be safely restarted after freeing. (PORT = net timer ID).

ACTION: If you suspect the timer has not been freed, another remote possibility is data corruption of the timer entry. If necessary, submit an SR.

CAUSE: An attempt was made to restart a net timer which had already been freed, or a timer which was in some sort of bad state. Since freed net timers can be immediately reallocated to anyone, they cannot be safely restarted after freeing. (PORT = net timer ID).

ACTION: If you suspect the timer has not been freed, another remote possibility is data corruption of the timer entry. If necessary, submit an SR.

CAUSE: An attempt was made to stop a net timer which had already been freed, or a timer which was in some sort of bad state. Since freed net timers can be immediately reallocated to anyone, they cannot be safely stopped after freeing. (PORT = net timer ID).

ACTION: If you suspect the timer has not been freed, another remote possibility is data corruption of the timer entry. If necessary, submit an SR.

CAUSE: This is supplemental logging to a previous Net Timers error. While getting ready to pop one or more net timers, there was a problem restarting the system timer associated with the timer list those net timers came from. Unless the list's system timer is restarted, no other net timers on that list will pop from now on.

ACTION: Review information regarding the previous errors, for possible hints as to the cause. If necessary, submit an SR, noting this and previous Net Timers errors.

CAUSE: While popping a net timer, a Transport timer trap- routine reported a nonzero status. In response, Net Timers has freed the timer entry which popped (PARM = status from trap routine, PORT = net timer ID).

ACTION: A part of the normal operation of Net Timers, this information is provided for Transport debugging purposes only. Any SR filed or action taken should address the problem of the failing trap routine, whose error status was reported in PARM. Operation of Net Timers should be unaffected.

CAUSE: Instead of a timeout or other supported message, the Net Timers port received a message which was not recognized or supported. Illegal messages cannot be acted upon; they are dropped and their message frames returned to the system (PORT = address of message frame, PARM = message descriptor).

ACTION: The operation of Net Timers should be unaffected. However, some other module on the system seems to be sending the wrong messages to the wrong places. Submit an SR, specifying the PARM value which was logged.

CAUSE: Reported by Hewlett-Packard internal tests only. This indicates resetting of a popping net timer yielded a status other than one indicating the timer was currently popping.

ACTION: This should never been seen outside the factory; submit an SR.

CAUSE: Reported by Hewlett-Packard internal tests only. This indicates a net timer trap- routine reported a false error, leading to release of the popped timer.

ACTION: This should never been seen outside the factory; submit an SR.

CAUSE: Reported by Hewlett-Packard internal tests only. This indicates freeing of a popping net timer yielded a status other than one indicating the timer was currently popping.

ACTION: This should never been seen outside the factory; submit an SR.

CAUSE: Reported by Hewlett-Packard internal tests only. This indicates freeing of a popping net timer, already previously freed, yielded a status other than one indicating the timer was not in use.

ACTION: This should never been seen outside the factory; submit an SR.

CAUSE: Reported by Hewlett-Packard internal tests only. This indicates resetting of a popping net timer, previously freed, yielded a status other than one indicating the timer was not in use.

ACTION: This should never been seen outside the factory; submit an SR.

CAUSE: While attempting to freeze a caller- specified statistics buffer into memory, the system routine Freeze reported an error. Freezing is required, to avoid page faults while Net Timers data structures are locked for the statistics report (PARM = status from Freeze).

ACTION: Net Timers will continue to run as normal, though no statistics data was obtained. Whichever utility was trying to read Net Timers information did not get it. Submit an SR.

CAUSE: A caller-specified statistics buffer was too small to hold the entire statistics block. Partial statistics- reads are not supported; to read statistics only, the caller must specify the exact length of the Net Timers statistics block.

ACTION: No statistics were returned. If this error was reported by a Hewlett-Packard-supplied program, submit an SR against that program.

CAUSE: A caller-specified statistics buffer was large enough to hold statistics, but not large enough to hold all the timer map data. Net Timers ran out of room trying to record a timer list head. Partial reads of timer- map data are not recommended.

ACTION: Statistics were returned, but the timer map data returned is incomplete; caller should have specified a larger buffer. If this error was reported by a Hewlett-Packard-supplied program, submit an SR against that program.

CAUSE: A caller-specified statistics buffer was large enough to hold statistics, but not large enough to hold all the timer map data. Net Timers ran out of room trying to record a timer interval. Partial reads of timer- map data are not recommended.

ACTION: Statistics were returned, but the timer map data returned is incomplete; caller should have specified a larger buffer. If this error was reported by a Hewlett-Packard-supplied program, submit an SR against that program.

CAUSE: A caller-specified statistics buffer was large enough to hold statistics, but not large enough to hold all the timer map data. Net Timers ran out of room trying to record a timer entry. Partial reads of timer- map data are not recommended.

ACTION: Statistics were returned, but the timer map data returned is incomplete; caller should have specified a larger buffer. If this error was reported by a Hewlett-Packard-supplied program, submit an SR against that program.

CAUSE: While attempting to unfreeze a caller- specified statistics buffer previously frozen into memory, the system routine Unfreeze reported an error. Unfreezing is necessary to avoid memory fragmentation over time (PARM = status from Unfreeze).

ACTION: Net Timers will continue to run as normal, though no statistics data was obtained, and some memory space may be lost until the next system startup. Submit an SR.

CAUSE: An attempt was made to perform a Net Timers control operation during Net Timers startup, or during or after shutdown. Control operations cannot be performed unless Net Timers is completely up.

ACTION: No control operation was performed. If this occurs during network shutdown, it can probably be ignored. If it occurs at any other time, submit an SR.

CAUSE: An unrecognized or unsupported function code was passed in a call to NETTMR_CONTROL. Only supported control functions can be acted on.

ACTION: The call was ignored; the operation of Net Timers should be unaffected. However, some other module on the system seems to be passing bad function codes. If necessary, submit an SR.

CAUSE: An attempt was made to shut Net Timers down before a successful startup, or during or after a shutdown. Multiple shutdowns are not allowed.

ACTION: No harm was done, however some other module on the system, possibly CP, seems to be confused. If necessary, submit an SR.

CAUSE: Net Timers is being shut down (PARM = Net Timers port data address).

ACTION: No action is required.

CAUSE: While releasing in use system timers during shutdown, the system routine Release_timer reported an error. All in use timers are normally deallocated during any shutdown of Net Timers (PORT = unfreed system timer ID, PARM = status from Release_timer).

ACTION: Net Timers will report the failure as a warning and continue with the shutdown. Unless there is corruption within system timers, this probably indicates a Net Timers bug. Another remote possibility is that the Net Timers port data has been corrupted from outside, so other, possibly serious problems may soon appear as well. In addition, one system timer may have been lost; if this continues, the system could eventually run out of system timers. Submit an SR.

CAUSE: While Net Timers was attempting to release its message port during shutdown, the system routine Purge_port reported an error. Purging the port is a normal part of resource cleanup during any Net Timers shutdown. Releasing port ownership during startup should have insured the port could be purged later, even if by a different process (PORT = Net Timers port ID, PARM = status from Purge_port.

ACTION: Net Timers will report the failure as a warning and continue with the shutdown. One message port may have been lost, but probably not any message frames; if this cont8inues, the system could eventually run out of ports. Submit an SR.

CAUSE: During shutdown, while Net Timers was attempting to unfreeze its previously frozen context memory (port data area or PDA), the system routine Unfreeze reported an error. Unfreezing the PDA is normally done during any Net Timers shutdown, in preparation for freeing the PDA object (PORT = port data address, PARM = status from Unfreeze).

ACTION: Net Timers will report the failure as a warning and continue with the shutdown. Some memory space may be lost until the next system startup, but unless this happens repeatedly, the system should continue to run. Submit an SR.

CAUSE: During shutdown, while Net Timers was attempting to release its context memory (port data area or PDA), the system routine Release_object reported an error. Freeing of the PDA is the final stage of resource cleanup by Net Timers during any shutdown (PORT = port data address, PARM = status from Release_object).

ACTION: Net Timers will report the failure as a warning and continue with the shutdown. Some memory space may be lost until the next system startup, but unless this happens repeatedly, the system should continue to run. Submit an SR.

CAUSE: While attempting to compute its configuration limits from the NETXPORT. GPROT. TCP node of the NMCONFIG. PUB. SYS file, the system routine Nmconfgetdata reported an error. Unable to read config data, a default configuration was selected as a backup just to get the network going (PARM = status from Nmconfgetdata).

ACTION: This failure was reported as a warning and Net Timers then continued with the startup, using a large default configuration. It is unlikely the default configuration will match your needs as well as a computed one, however your network may still run correctly. Check to make sure you have configured the TCP protocol into your network, specifying some maximum number of connections. Also check to make sure the NMCONFIG file is available and not being accessed at the current time. If necessary, submit an SR, specifying the PARM value from the logging message.

CAUSE: An attempt was made to start Net Timers up during or after a successful startup, but before a complete shutdown. Multiple startups are not allowed.

ACTION: No harm was done, however some other module on the system, possibly CP, seems to be confused. If necessary, submit an SR.

CAUSE: After Net Timers computed its configuration during startup, an attempt to allocate context memory (port data area or PDA) using the system routine Create_object, failed. The PDA is where all data structures needed to operate Net Timers must reside (PARM = status from Create_object).

ACTION: Depending on the PARM value, your system load, and your current configuration, it is possible not enough free memory exists on your system to support the number of timers required by maximum the number of TCP connections currently configured in NMMGR. If this seems likely, you might try starting your network sooner after system startup, or possibly reducing the number of TCP connections configured. If necessary, submit an SR.

CAUSE: After Net Timers successfully allocated an object for its context memory (port data area or PDA), the system routine Freeze reported an error tying to make the entire PDA present in physical memory. The entire PDA must be frozen so it can be accessed from the Interrupt Control Stack by Transport, and during timer pops (PARM = status from Freeze).

ACTION: Depending on the PARM value, your system load, and your current configuration, it is possible too much resident memory is already in use on your system to support the number of timers required by maximum the number of TCP connections currently configured in NMMGR, and thus the required PDA size. If this seems likely, you might try starting your network sooner after system startup, or possibly reducing the number of TCP connections configured. If necessary, submit an SR.

CAUSE: During startup, when attempting to load the address of an NL.PUB.SYS procedure to service the Net Timers port, the system routine Hpgetsysplabel reported an error trying to locate the system routine nettmr_ics_server. This procedure is required so the system will know who to call to process timeout messages arriving at the port (PARM = status from Hpgetsysplabel).

ACTION: Net Timers failed to start; instead it cleaned up after itself and reported an error. If you just installed a new release of the Transport subsystem, possibly the installation was incorrect, incomplete, or corrupted. There is also a remote possibility the system has run out of room for dynamically- loaded procedures. If necessary, submit an SR.

CAUSE: During startup, while attempting to create a message port to receive system-timer timeout messages, the system routine Create_port failed. The port is needed as a destination for timeout messages for the system timers which drive Net Timers (PARM = status from Create_port).

ACTION: Net Timers failed to start; instead it cleaned up after itself and reported an error. If your system is heavily loaded, it is possible the system has run out of ports; if you think this is the case, you might try reducing system load, however a significant reduction would be needed to allow all of Transport to start up. If necessary, submit an SR.

CAUSE: During startup, after successful creation of a port for Net Timers, the system routine Port_control reported an error trying to release CP's ownership of the port. Port ownership must be released in case a different process deletes the Net Timers port later (PARM = status from Port_control).

ACTION: Net Timers failed to start; instead it cleaned up after itself and reported an error. Submit an SR.

CAUSE: During startup initialization of the Net Timers context memory (port data area or PDA), the system routine Cb_init reported an error trying to clear the Net Timers semaphore. The semaphore is required because multiple callers can access the same timer list, and that list can pop at any time (PARM = status from Cb_init).

ACTION: Net Timers failed to start; instead it cleaned up after itself and reported an error. Submit an SR.

CAUSE: Net Timers is being started up (PARM = Net Timers port data address).

ACTION: This logging is provided for information purposes only. No action is required.