Displayed Information on the Alert Server (SAP Library

Displayed Information on the Alert Server

Use

The TREX alert server provides information on three tab pages.

● Summary: Displays a summary of the system status

● Current: Displays the status of all checks in the last check run

● History: Displays the status of all checks in the previous check runs

Integration

The tab pages mentioned above appear in the TREX admin tool (stand-alone) and in the TREX admin tool in the SAP system.

Features

Summary Tab Page

The Summary tab page displays the current system status. TREX determines this status using the results that were gathered during the last check run. If this contains too many errors, the status yellow or red is displayed.

The messages on this tab page are displayed in a fixed sequence. If more than one message is displayed at a time, you work through the instructions from top to bottom.

Information on the Summary Tab Page

Column	Description
Status	Displays the current overall status ● Green: There are no errors. ● Yellow: There are errors. You should solve these errors. ● Red: There are serious errors. Solve these errors promptly because the system is not working correctly.
Number of Details	Specifies the number of detailed information items available
Action	Instructions for administrators on solving the problems causing this status
Originating Checks	Specifies the checks that have led to this status

Current Tab Page

The Current tab page displays the results that were gathered during the last check run.

To sort the table content by the content of a column, click the header of the column in question in the TREX admin tool (stand-alone). If you are using the admin tool in the SAP system, you also choose the Sort function. An arrow shows whether the column is sorted in ascending or descending order.

Information on the Current Tab Page

Column	Description
Status	Displays the current status ● Green: There are no errors. ● Yellow: There are errors. You should solve these errors. ● Red: There are serious errors. Solve these errors promptly because the system is not working correctly.
Check Name	Name of the check
Check Set	Name of the check set
Age	Specifies the time in seconds since the check took place
Duration	Duration of the check in milliseconds
Check State	Status of the check: running: The check is in progress, the results of the previous check run are displayed ready: The check has been run, the results of the last check run are displayed
Number of Details	Specifies the number of detailed information items available
Check Description	Description of the check
Action	Instructions for administrators on solving the problems causing this status

History Tab Page

Use the information on the History tab page to analyze the results of previous checks.

Information on the History Tab Page

Column	Description
Status	Displays the current status ● Green: There are no errors. ● Yellow: There are errors. You should solve these errors. ● Red: There are serious errors. Solve these errors promptly because the system is not working correctly.
Check Name	Name of the check
Check Set	Name of the check set that contains the check
Execution Time Stamp	Timestamp of the last run
Duration	Duration of the check in milliseconds
Number of Details	Specifies the number of detailed information items available
Check Description	Description of the check
Action	Instructions for administrators on solving the problems causing this status

Alert Server Check Descriptions

The table below contains an alphabetical list of the available alert server checks. You carry out different activities depending on whether the status is This graphic is explained in the accompanying text (yellow) or (red). There are some checks that cause a message to be issued but that do not require any activity.

Explanation of check sets:

● hpa_1 = BIA scenario (no queue server, distributed landscape)

● hpa_2 = BIA scenario (no queue server, distributed landscape)

● standard_1 = Scenario for the portal, KM, and A1S config

● standard_2 = Scenario for the portal, KM, and A1S config

Note

The hpa_1 and hpa_2 check sets and the standard_1 and standard_2 check sets are executed at different times.

● D_A_R = 'Detect_And_Repair': Check set with checks that run every day

Check Name	Message	Activity	Check set
alertserver_active_master	The alert server shown in the screen area below is now the active master.	No action is required.	Internal check
alertserver_start	The Alert Server has been started.	No action is required.	Internal check
alertserver_stop	The alert server has been shut down.	No action is required.	Internal check
ask_trex_expert	An unknown error occurred.	Contact a TREX expert. Open a customer message.	Internal check
check_not_active_because_mem_watcher_is_on	Check was set to not active because the mem watcher is aware of it.	No action is required.	Internal check
check_topology_index_garbage	Returns a list of indexes which exist with wrong information in topology.	Open a customer message with information from the results list. See SAP note 1101958 for details.	standard_1 hpa_1
cpu_usage	Returns CPU usage.	If necessary, check your hardware sizing. Threshold values: ● Warning/yellow : 70% CPU usage ● Error/red : 90% CPU usage	standard_1 hpa_1
detect_and_remove_obsolete_filter _trace_files	Checks the existence of obsolete TrexFilter.trc files and removes them if found.	The files below are trace files from the Inxight filters that will not be used anymore. They have been deleted. No action is required.	standard_2
detect_core_files	Returns a list of core files.	Open a customer message. See SAP Note 1101958.	standard_1 hpa_1
delta_index_correct	Checks the correct delta index for indexes in namespace 1fs	Delta index (indexes) has (have) been corrected. No action is required.	D_A_R
delta_index_status	Returns all indexes whose delta index is too big (more than <threshold_value> documents).	Merge the delta indexes into the main index.	standard_1 hpa_1
disk_usage	Returns any host for which the available disk space is under a certain limit.	Increase the available disk space. Threshold values: ● Warning/yellow : 80% disk usage ● Error/red : 90% disk usage	standard_1 hpa_1
failed_replication	Returns all indexes for which replication failed.	Trigger index replication.	standard_1
filer_writeability	Checks if a filer is permanently writable (check normally disabled). WARNING: This check fires huge files against the filer. Means high TREX performance decrease!	Resolve filer problems. Disable filer_writability check if no longer needed.	standard_1 hpa_1
host_consistency	Returns a list of hosts and the compile time (version).	If necessary, make the different versions of the TREX hosts in the TREX landscape consistent. All hosts in your landscape must have the same release and version.	standard_1 hpa_1
inactive_backup_server	Returns the host and port where the TREX backup server has been set up but does not run.	Recreate the affected indexes. See More Details column in results list.	D_A_R
inactive_http_server	Returns all inactive http servers.	Check that TREX is up and running in the system in question. If not, start TREX; otherwise, check that the HTTP server is running in the system in question. If not, start the HTTP server.	standard_1
inactive_index _servers	Returns a list of inactive index servers.	Check that the index server is up and running in the system in question. If not, start the index server.	standard_1 hpa_1
inactive_name_servers	Returns a list of inactive name servers.	Check that TREX is up and running in the system in question. If not, start TREX; otherwise, check that the name server is running in the system in question. If not, start the name server.	standard_1 hpa_1
Inactive_preprocessors	Returns a list of inactive preprocessors.	Check that TREX is up and running in the system in question. If not, start TREX; otherwise, check that the preprocessor is running in the system in question. If not, start the preprocessor.	standard_1 hpa_1
inactive_queue_servers	Returns a list of inactive queue servers.	Check that TREX is up and running in the system in question. If not, start TREX; otherwise, check that the queue server is running in the system in question. If not, start the queue server.	standard_1
inconsistent_logical_indexes	Returns a list of inconsistent logical indexes.	Check whether the physical index (or indexes) of the listed logical indexes have the status 'corrupted'. If so, recreate the physical index (or indexes) using the application replication mechanism. If the physical index (or indexes) has (have) the status 'pending', simply wait. If not, check that TREX is up and running in the system in question. If not, start TREX.	standard_1 hpa_1
index_data_inconsistencies	Returns a list of indexes with inconsistent data. This check may cause heavy load on the system and should only run if needed.	Recreate the affected indexes. See the More Details column in the results list.	D_A_R
index_server_round_robin	Returns any indexes with a round robin problem.	Open a customer message.	standard_1
index_status	Returns any indexes with an error state.	Check the index status. Depends on the application using TREX. Standard recommendation is to re-index the index in question.	standard_1 hpa_1
internal_error	An internal error occurred while executing the check	Open a customer message.
invalid_cron_format	The schedule format is not correct.	Fix the schedule format of the relevant check set(s).	standard_1
invalid_master_snapshots	Returns all master indexes whose snapshots and sync points differ (because the replication was not started)	Trigger replication.	standard_1
invalid_slave_snapshots_1	Returns all slaves with snapshots older than the master snapshots (because the replication was started but the import did not work)	Trigger index replication.	standard_1
invalid_slave_snapshots_2	Returns all slaves with snapshots older than the master sync points (because the replication was not started)	Trigger index replication	standard_1
lonesome_idx	Checks the existence of physical indexes which are not longer part of a cube.	Check whether or not the index is in use. If not you may delete the index.	standard_1
mailer_error	An error occurred while sending a notification mail	Check your alert server mail configuration. Hints can be found in the message text.
master_indexes_running_at_backup	Returns a list of indexes running on the backup server	Find out why the master queue server is not running.	standard_1
memory_usage	Returns memory usage	If necessary, check your hardware sizing. Threshold values: ● Warning/Yellow : 75% memory usage ● Error/Red : 90% memory usage	standard_1 hpa_1
merge_deltas	In non-BIA systems the merge_delta parameter has to be switched on.	merge_deltas	standard_2
meta_model_cycle	Returns a list of cycles in the meta model (joins). Meta model cycles cost performance. A cycle exists i.e. when index A is joined to index B, B to C and index C is joined back to index A (with the same semantic as index A mentioned first) again.	The affected meta model view is provided by SAP. Open a customer message or rework your meta model. See SAP note 1101958.	D_A_R
multiple_sids_at_bia	Checks the existence of multiple SAPSIDs connected to a BIA system.	The connection of multiple SAPSIDs to a BIA system is not recommended. Use only one BIA for each SAPSID.	hpa_1
network	Returns hosts without network connection.	Check the hardware.	standard_1 hpa_1
no_master_index_server	Returns a list of indexes without active master or backup.	Check that the index server is up and running in the system in question. If not, start the index server.	standard_1 hpa_1
no_master_queue_server	Returns a list of queues without active master or backup.	Check that TREX is up and running in the system in question. If not start TREX; otherwise check that the queue server is running in the system in question. If not, start the queue server.	standard_1
no_slave_index_server	Returns a list of indexes without an assigned slave server.	Check that the index server is up and running in the system in question. If not, start the index server.	standard_1
open_handles	Returns the open handles.	Check your kernel limits, decrease the number of loaded indexes, or check your hardware sizing. Threshold values: ● Warning/Yellow : 1000 open handles ● Error/Red : 1500 open handles	standard_1 hpa_1
pending_index_server_requests	Returns all index servers with <threshold_value> or more pending requests.	Check the configuration of your landscape.	standard_1 hpa_1
pending_preprocessor_requests	Returns all preprocessors with <threshold_value> or more pending requests.	Check the configuration of your landscape.	standard_1
pending_queue_server_requests	Returns all queue servers with <threshold_value> or more pending requests.	Check the configuration of your landscape.	standard_1
ping	Returns the ping time.	No action is required.
Preprocessor_round_robin	Returns any preprocessor type with a round robin problem.	Open a customer message.	standard_1
psp_appserver_check	Check if PSP application server is running.	Check why your PSP application server has not started.
queue_failed_states	Returns all queues with documents in an error state.	Reset the documents with an error state.	standard_1
queue_server_round_robin	Returns all queues with a round robin problem.	Open a customer message.	standard_1
queue_status	Returns all queues with an error state.	Check the queue status.	standard_1
queues_running_at_backup	Returns a list of queues running on the backup server	Check why the master queue server is not running.	standard_1
reorg	Returns yes if the reorganization of the landscape is necessary.	Reorganize the TREX landscape.	standard_2 hpa_2
reset_landscape	“Reset Landscape” was executed.	The TREX landscape has been reset. Delete all indexes and reboot all hosts.
restarted_services	Returns all restarted services	If restart was not carried out manually, open a customer message.	standard_1 hpa_1
rfc_connection	Checks RFC server configuration and tries to repair if necessary	Possible cause/action: -- Manual repair -- RFC configuration is set to 'manual', so the check should not be active. Check will be disabled. -- RFC configuration is set to 'disabled', so the check should not be active. Check will be disabled. -- Connectivity data missing: If RFC is to be used, enter connectivity information. Otherwise, switch RFC off. -- TREXDaemon.ini incomplete or corrupted. Manual repair required. -- TREXRfcServer.ini incomplete or corrupted. Manual repair required. -- RFC Connection Error. Action see details. -- ABAP Error. Manual repair required. --RFC Error. Check details. -- Automatic repair performed. -- Application(s) on other host(s) use the same RFC destination. Manual repair required. -- Automatic repair not possible. Manual repair required.	standard_2 hpa_2
shared_memory	Returns shared memory usage.	Possible actions: -- Increase shared memory size. -- Set shared memory to same size on all hosts. See SAP notes 991567 and 997772 for details. Threshold values: Warning/Yellow : 75% usage ● Error/Red : 85% usage	standard_1 hpa_1
shared_memory_use	Returns list of hosts which do not use shared memory.	Check why TREX hosts do not use shared memory. See SAP Note 1059260.	standard_1 hpa_1
statistic_info	Returns statistic information.	If necessary, check your hardware sizing.
trace_entries	Returns critical trace file entries.	Open a customer message and specify the trace results. See SAP Note 1101958 for details.	standard_1 hpa_1
trace_file_size	Returns a list of huge trace files.	Check trace file directories - and delete/move the huge files. ● Warning/Yellow : 50 MB trace file size ● Error/Red : 2000 MB trace file size	standard_1 hpa_1
unassigned_logical_indexes	Returns a list of unassigned logical indexes.	Check that the index server is up and running on the system in question. If not, start the index server.	standard_1
unassigned_master_indexes	Returns a list of unassigned master indexes.	Check whether or not the index has the status 'corrupted'. If necessart, recreate the index using the application replication mechanism. If the index has the state 'pending', simply wait. Otherwise, check that TREX is up and running in the system in question. If not, restart TREX.	standard_1 hpa_1
Unassigned_queues	Returns a list of unassigned queues.	Restart the queue server	standard_1
Unassigned_slave_indexes	Returns all active slave index servers with an unassigned slave index.	Check whether or not the index has the status 'corrupted'. If so, recreate the index using the application replication mechanism. If the index has the status 'pending', simply wait. Otherwise, check whether TREX is up and running in the system in question. If not, start TREX.	standard_1
unloads	Returns any index that was unloaded from memory.	Either increase the memory on the host in question, carry out a reorganziation (distributed systems only), or decrease the number of loaded indexes. Threshold values: ● Warning/Yellow : 20 loaded indexes ● Error/Red : 30 loaded indexes	standard_1 hpa_1
version_info	Returns version info for each host.	If necessary, make the different versions of the TREX hosts in the landscape consistent.	standard_1 hpa_1

Activities

This graphic is explained in the accompanying text

In the production system, you do not have to have the Alert window open constantly and monitor it, because you can have the system notify you by e-mail in the event of critical system statuses.

To refresh the display of the tab pages, choose Refresh.

If a check returns a yellow or red traffic light, you should read and work through the instructions in the Action column.

To call up the details of a check, double-click the associated row. The detailed information is displayed in the TREX admin tool (stand-alone) below the table. In the SAP system, the detailed information is opened in a new window.