Table of Contents

Database Tables

Viewing Report Tables
Database Table Quick Reference

Understanding the Reporter Daemons

Snapshot Daemon

Configuration Control Files

Extraction Daemon
Rollup & Aging Daemon
Database Server Daemon
Check Daemon
IP Ping Daemon
Daemon Logs

Understanding Daemon Log and Control Files
Database Table Schemata
Specifying Report Parameters

Reviewing Report
Resource Information

You can use a number of tools within the TrafficDirector application to work with the database that contains all information for Trend Reporting (NSTREND_DB), the statistics contained in each database table, and information on how the TrafficDirector daemons work to retrieve and process the information you request when you generate reports.

The following sections describe the tools you can use to review report resource information:

• Database Tables

• Understanding the Reporter Daemons

• Understanding Daemon Log and Control Files

• Database Table Schemata

• Specifying Report Parameters

Database Tables

When you use reporting features, the queries you submit to the database can span multiple agents and domains. The following eight types of report tables are available:

• Protocol statistics

• Ethernet segment statistics

• Token Ring segment statistics

• Fiber Distributed Data Interface (FDDI) segment statistics

• WAN segment statistics

• Host statistics

• Conversation statistics

• Round Trip Delay statistics

Table 34-1 shows how each report table type can be classified into one or more categories.

• Viewing Report Tables

• Database Table Quick Reference

Viewing Report Tables

You can look at any of the preceding report tables for the types of content shown in Table 34-2.

Table 34-2: Report Content

This Report Type...	Contains This Type of Information
Snapshot	Raw snapshot data; used as a resource to produce usable statistics for the detail and daily tables. The snapshot daemon creates each new row of data in the snapshot table. The database extraction daemon uses each available pair of rows in the snapshot table to create a corresponding row in the detail table. After the database extraction daemon has used a pair of snapshot table rows, the older row is discarded, and the second (newer) row becomes the older row after new data has been received and logged.
This Report Type...	Contains This Type of Information
Details	Highly detailed data ranging from 1 minute to 24 hours, depending on how you define logging and polling parameters. Direct queries to detail tables when you want reports to reflect the details of the statistics collected daily from a SwitchProbe device.
Daily	Utilization data averaged for the day. For all other statistics, daily tables contain delta values that represent counters of units between a specific start and end time. What you get from the daily table is one data point on a graph that represents that one day out of n number of days.

Database Table Quick Reference

For a complete list of the statistics that are logged in each database table, see "Database Table Schemata."

Table 34-3 lists all tables in the database.

Table 34-3: Database Tables

Table Type	Table Name--- On UNIX On Windows	Contents/ Bytes per Row
Protocol Details	protocoldt_<yymmdd> protocol_detail	Media-independent protocol statistics details 165
Protocol Summary	protocol_summary	Media-independent protocol daily statistics 165
Ethernet Summary (*)	seg_et_detail	Ethernet segment statistics details 280
Ethernet Details	segetdt_<mmddyy> set_et_summary	Ethernet segment daily statistics 280
WAN Details	segwandt_<yymmdd> seg_wan_detail	WAN segment statistics details 280
WAN Summary (*)	seg_wan_summary	WAN segment daily statistics 280
FDDI Details	segfddit_<yymmdd> seg_fddi_detail	FDDI segment statistics details 280
FDDI Summary (*)	seg_fddi_summary	FDDI segment daily statistics 280
Token Ring Details	segtrdt_<yymmdd> seg_tr_detail	Token Ring segment statistics details 469
Token Ring Summary (*)	seg_tr_summary	Token Ring segment daily statistics 469
Host Details	hostdt_<yymmdd> host_detail	Host statistics details 225
Host Summary (*)	host_summary	Host statistics daily statistics 225
Conversation Details	convdt_<yymmdd> conv_detail	Conversation statistics details 201
Conversation Summary (*)	conv_summary	Conversation daily statistics 201
IP Ping Details	ippingdt_<yymmdd> ipping_detail	Round Trip Delay statistics detail 145
IP Ping Summary (*)	ip_ping_summary	Round Trip Delay daily statistics 145
ART Details	artdt_<yymmdd> art_details	Application Response Time statistics details 287
ART Summary (*)	art_summary	Application Response Time daily Statistics 287
ART Server Details	artsummdt_<yymmdd> artsumm_detail	Application Response Time server statistics details 287
ART Server Summary(*)	artsumm_summary	Application Response Time Server daily statistics 280
SNMP Get Details	snmpgetdt_<yymmdd> snmp_get_detail	Proxy SNMP statistics details 280
SNMP Get Summary (*)	snmp_get_summary	Proxy SNMP daily statistics 280
Network Conversation Matrix TopN Details	nltopncdt_<yymmdd> nltopnc_detail	Network Conversation Matrix TopN statistics details 207
Network Conversation Matrix TopN Summary (*)	nltopnc_summary	Network Conversation Matrix TopN daily statistics 207
Application Conversation TopN Details	altopncdt_<yymmdd> altopnc_detail	Application Conversation Matrix TopN statistics details 222
Application Conversation TopN Summary (*)	altopnc_summary	Application Conversation Matrix TopN daily statistics 222
(*) If Peak-Business-Hours (PBH) (hour-range) is defined other than 00:00 to 24:00 in dbupdate.cfg, the Daily SQL tables represent the (N) hours of PBH. The PBH hour range = (hour range end time - hour range start time) and may be any range from 1 to 24 hours.

Understanding the Reporter Daemons

Daemons are processes that run in the background and are disconnected from a process group and terminal. The TrafficDirector application uses specific daemons to create and update tables in the report database. Daemons work with related configuration files that rule how and when they are called to perform their roles.

The daemons listed in Table 34-4 work with the SQL database and Trend Reporter.

Table 34-4: Daemons That Work With SQL Database
and Trend Reporter

This Daemon...	Performs This Function
Snapshot daemon (dbsnapd)	Retrieves snapshots of raw statistics from agents and stores this information in a snapshot table.
Extraction daemon (dbextrad)	Started by the dbchkd daemon and migrates data records from binary files to a SQL database (NT platforms only). Retrieves information from the snapshot table, creates new rows in the details table, and deletes the oldest member of each pair of snapshot rows.
Rollup and aging daemon (dbrolld)	Performs two separate but related actions---Retrieves information from the details table to create new rows in the summary table, and deletes obsolete information from both the detail and summary tables.
SQL server daemon (msqld)	Started by the dbchkd daemon and migrates data records from binary files to the embedded SQL database. Handles reporting queries. Listens for and processes SQL queries.
This Daemon...	Performs This Function
Check daemon (dbchkd)	Started by the TrafficDirector application, this daemon starts the following daemons: Database server daemon (msqld) Trap daemon (dvtrapd) Snapshot daemons (dbsnapd) IP ping snapshot daemon (dbsnpres) Extraction daemon (dbextrad) Rollup daemon (dbrolld) At regular intervals, dbchkd verifies that all child daemons are alive and, if necessary, restarts them. Also initializes the database, creates database tables and the directory structure for storing the binary files ($NSHOME/db/snapshot, detail and summary). For embedded SQL configurations, summary tables are created once; however, detail tables are created at the start of every day. For Microsoft SQL configurations (NT platforms only), dbchkd creates one set of detail and summary tables.
IP Ping daemon (dbsnpres)	Started by the dbchkd daemon, this daemon polls network agents and logs information to the snapshot, detail, and summary IP ping binary files. Also responsible for polling and logging proxy SNMP data.
Autoreporter daemon (autorptd)	Started by the dbchkd daemon, sequentially runs all reports configured in the autorpt.cfg file. This daemon starts up daily at 3:00 a.m. Can also be started by using the ezreport utility with the -period or -days parameter.
Trap daemon (dvtrapd)	Started by the dbchkd daemon, this daemon listens for and logs SNMP trap messages. Aso receives and logs important informational messages from all daemons. Logs messages to the YYMMDD.fdm file to make these messages available to the Alert Monitor.

• Snapshot Daemon

• Extraction Daemon

• Rollup & Aging Daemon

• Database Server Daemon

• Check Daemon

• IP Ping Daemon

• Daemon Logs

Snapshot Daemon

The dbsnapd daemons are started to perform logging for agents, Frame Relay agents, and switches, based on the current list of names in the *.*lg files created by the dbchkd daemon. Two dbsnapd daemons are started for agents/Frame Relay agents---one for logging STATS and one for logging Hosts and Convs (even if both are not being logged). One dbsnapd is started for logging stats for all ports on one switch.

If no log configuration is present for Host and Conversation, the corresponding dbsnapd terminates. dbchkd restarts this daemon every 15 minutes to retest for configuration changes.

Each dbsnapd daemon opens the *.*lg file passed to it as an argument by the dbchkd daemon and logs either Stats or Hosts/Convs to snapshot binary files.

dbsnapd periodically checks to see if records from the detail binary files have been migrated to the database detail tables by msqld or dbextrad. If a backup of detail binary files is detected, the following error message displays.

"*** WARNING!!! SQL Server lagging behind skipped * logging for this interval!" (For mSQL mode)
 
"*** WARNING!!! Extraction daemon lagging behind skipped * logging for this interval!" (For Microsoft SQL mode)

See Appendix H, "Error Messages" for detailed information on error messages.

Each snapshot binary file name includes: the device name, domain, date/time (YYMMDDHHMM) and logging interval. For example: et-101.RMON.9710031030.00:15

These files are stored according to the following directory structure:

NSHOME/db/snapshot:	/protocol:	
/altopn_c
/art
/artsumm
/conv
/host
/ip_ping
/nltopn_c
/protocol
/seg_et
/seg_fddi
/seg_tr
/seg_wan
/snmp_get

The snapshot daemon is also responsible for calculating the change (deltas) between the counters on file from the device, and the just polled counters, and logging the detail records to detail binary files.

Each detail binary file name includes: the device name, domain, date/time (YYMMDDHHMM) and logging interval. For example: et-101.RMON.9710031030.00:15

Delta values are stored in the detail binary file according to the following directory structure:

NSHOME/db/detail: 	/protocol:	
/altopn_c
/art
/artsumm
/conv
/host
/ip_ping
/nltopn_c
/protocol
/seg_et
/seg_fddi
/seg_tr
/seg_wan
/snmp_get

The detail records from the detail binary files are migrated to the database by the msqld daemon (embedded SQL configuration), or dbextrad daemon (Microsoft SQL configuration).

As detail binary files are created, dbsnapd aggregates the individual files into a summary binary record so that at the end of the day, it has a complete summary record for the day.

Continuing with the previous example, the first detail binary file created (D1) also becomes S1, (at this time the summary file S1 is the same as D1). At the next logging interval, a new detail binary file is created (D2), the summary file S1 is read, and the contents of D2 are added to S1. The summary file now contains D1 + D2. Continuing this process throughout the day, at the end of the day, S1 will contain the complete summary record.

The configuration setting for PBH, in dbupdate.cfg, affects the accumulation of summary data. Data is summarized only within the limits of Day Range start-end.

Summary data is stored in the summary binary file according to the following directory structure:

NSHOME/db/summary:	/protocol:	
/altopn_c
/art
/artsumm
/conv
/host
/ip_ping
/nltopn_c
/protocol
/seg_et
/seg_fddi
/seg_tr
/seg_wan
/snmp_get
 

Summary records from the summary binary files are migrated to the database, every
24 hours at 1:00 a.m., by the msqld daemon (embedded SQL configuration), or dbextrad daemon (Microsoft SQL configuration) during the rollup activity.

Configuration Control Files

When you use Configuration Manager to install properties files that include logging parameters on agents, switches, or Frame Relay agents the following files are created and saved to the $NSHOME/usr directory:

• <agent name>.alg---Agent logging configuration file

• <switch name>.slg---Switch logging configuration file

• <Frame Relay agent>.flg---Frame Relay agent logging configuration file

The dbchkd daemon makes a list of all the *.*lg files. For each name on the list, the dbchkd daemon starts dbsnapd daemons. Each file contains the agent and domain logging instructions and intervals that the dbsnapd daemons use to poll network agent data. Each line in a *.*lg file contains an agent name, domain name, and the intervals used for logging statistics, host, conversation, and ART information.

Extraction Daemon

The dbextrad daemon is used only on Windows NT platforms (Microsoft SQL configurations) and does the following:

• Migrates the records from the detail binary files to the database detail tables by periodically checking for the readiness of detail binary files; deletes the detail binary files once data is migrated.

• Daily at 1:00 a.m., migrates records from the summary binary files to the database summary tables; deletes the summary binary files once the data is migrated.

• Writes status messages reflecting the activity performed by dbextrad to a log file called $NSHOME/log.

Rollup & Aging Daemon

The dbchkd daemon starts the Aging daemon, dbrolld, daily at 2:00 a.m. to delete the rows in the database based on aging parameters configured in the $NSHOME/usr/dbupdate.cfg file. This ASCII configuration file contains parameters that determine how long data is stored in the detail and summary database tables before dbrolld deletes the obsolete rows in the database. You can change information in this configuration file by using the Config Rollup application within the TrafficDirector Admin level.

In the dbupdate.cfg file, you will find one line of information (parameters shown in number of days) for each type of table---protocol, segment, host, conversation, IP-Ping, Proxy-SNMP, ALL-NL, ALL-AL, Application Response Time, Application Response Time Server.

The dbupdate configuration file contains the following parameters for aging:

• Detail Aging---Shows the number of days that information ages before the Aging daemon determines it to be obsolete and deletes it from the details table. Detail aging must be greater than one day. The default value for protocol and segment statistics from the detail table is 31 days; for host, conversation, IP Ping, Proxy SNMP, ALL NL, ALL AL, ART, and ART Server, it is 7 days.

• Summary Aging---Shows the number of days that information ages before the Aging daemon determines it to be obsolete and deletes it from the daily table. Daily aging must be greater than one day. The default value for protocol and segment statistics from the Summary (Daily) table is 366 days; for host, conversation, IP Ping, Proxy SNMP, ALL NL, ALL AL, ART, and ART Server, it is 31 days.

Database Server Daemon

The msqld daemon is used with embedded SQL configurations for logging to an SQL database. In addition to satisfying reporting application requests for data, the msqld daemon performs the following tasks:

• Migrates the records from the detail binary files to the database detail tables by periodically checking for the readiness of detail binary files; deletes the detail binary files once data is migrated.

• Daily at 1:00 a.m., migrates records from the summary binary files to the database summary tables; deletes the summary binary files once the data is migrated.

• Writes status messages reflecting the activity performed by msqld to the msqld.log file.

• Completes a backup of msqld.log file at midnight and creates a backup file called msqld_day.log. For example, on Wednesday at midnight, the backup file created is msqld_Wed.log. Msqld_day.log data is lost after seven days.

Check Daemon

The TrafficDirector application is responsible for starting the check daemon (dbchkd). Once started, the msqld or dbextrad, dvtrapd, dbsnpres, and dbsnapd daemons are started. Every fifteen minutes, the TrafficDirector application checks to see if dbchkd is running, using a control file mechanism.

The primary task of dbchkd is to start snapshot daemons (dbsnapd) for each agent, switch, or Frame Relay agent for which logging has been configured. dbchkd makes a list of all logging configuration files (*.alg, *.flg, *.slg). For each name in the list, two daemons are started---one to do stats logging, and one to do host and conversation logging.

For example, the dbchkd daemon starts the two snapshot daemons for agent et-101 as follows:

$ dbsnapd STATS et-101.alg (daemon to do stats logging)
$ dbsnapd HOST_CONV et-101.alg (daemon to do host and conversation logging)

Every thirty seconds, dbchkd awakes and writes its process id (pid) to the dbchk.ctl file. Every minute, it makes a list of the *.*lg file names and compares the list to the previous list of *.*lg file names. From these two lists, a new comprehensive list of *.*lg file names is compiled.

• For file names not on the new list, the snapshot daemons are killed.

• For new file names on the list, snapshot daemons are started.

• For names on both lists, dbchkd checks to see if the current snapshot daemons are alive. If the snapshot daemons are not alive, new daemons are started.

The check daemon checks every five minutes for a disk full condition. There is a default threshold of 10 Mb against which the disk full check is made. If the disk on which logging is being done reaches the default threshold, all logging daemons are automatically stopped. You configure this threshold (min_disk_space) in the sqldb.dvp file ($NSHOME/usr), and you can modify it using a text editor.

If the disk full condition occurs, the user is notified by the Alert Monitor and $NSHOME/log/daemon-MMDD.log file, by the following error message:

"WARNING!!! No disk space to continue logging. Shutting down logging daemons.

The disk full check is performed on both embedded SQL and Microsoft SQL configurations. However, since SQL might be on a remote machine and the SQL database size is preconfigured, the disk full feature might not be applicable to your installation.

The check daemon also checks every five minutes to see if the msqld daemon, ip ping daemon (dbsnpres), and dbextrad (Microsoft SQL configuration only) daemons are alive. If any of these daemons are not alive, they are restarted.

The check daemon starts the aging daemon (dbrolld) daily at 2:00 a.m. and the autoreporter daemon (autorptd) daily at 3:00 a.m.

IP Ping Daemon

The IP ping daemon, dbsnpres, goes through the *.*lg files, and checks to see if ip_ping logging is configured. If it is configured, it polls network agents and logs IP ping data to the binary files. It works the same way as the snapshot daemons (dbsnapd). The IP ping daemon also polls and logs SNMP proxy data to the binary files.

Only one dbsnpres daemon is used to perform IP-Ping and Proxy-SNMP polling and logging activity for all configured agents.

Daemon Logs

All daemon messages, regardless of the type of daemons, (with the exception of msqld) are stored in $NSHOME/log/daemon-mmdd.log, where mmdd indicates the month and day the log was created.

Important (fatal) daemon messages are delivered to the Alert Monitor from dvtrapd.

For more information about configuring logging, and changing default values, see Appendix B, "Maintaining the TrafficDirector Environment."

Understanding Daemon Log and Control Files

All Trend Reporter true daemons (Snapshot, Extraction, and Rollup & Aging) record their activities by writing this information into log files. Control files are special files that contain code for the keepalive mechanism that helps the TrafficDirector application ensure that only one copy of each daemon is running at a given time.

Table 34-5: Daemon Log Files

Daemon	Associated Logfile	Associated Control File
dbsnapd	dbsnap.log	dbsnap.ctl
dbextrad	dbextra.log	dbextra.ctl
dbrolld	dbroll.log	No control file is created

Database Table Schemata

This section the table schemata used in the database. Tables in this section include:

• Table 34-6, Protocol Detail and Summary Tables

• Table 34-7, Segment Detail and Summary Tables (Ethernet-Specific)

• Table 34-8, Segment Detail and Summary Tables (WAN-Specific)

• Table 34-9, Segment Detail and Summary Tables (FDDI-Specific)

• Table 34-10, Segment Detail and Summary Table (Token Ring-Specific)

• Table 34-11, Host Detail and Summary Tables

• Table 34-12, Conversation Detail and Summary Tables

• Table 34-13, IP Ping Detail and Summary Tables

• Table 34-14, Application Response Time (ART) and Application Response Time (ART) Server Detail and Summary Tables

• Table 34-15, SNMP Get Detail and Summary Tables

• Table 34-16, Network Conversation Matrix TopN (nltopnc) Detail and Summary Tables

• Table 34-17, Application Conversation Matrix TopN (altopnc) Detail and Summary Tables

Table 34-6: Protocol Detail and Summary Tables

Field	Type	Description
Agent	char(26)	Agent name.
Domain	char(15)	Domain name.
StartTime	int	Start time.
EndTime	int	End time.
StartTimeText	char(15)	Start time (text format).
EndTimeText	char(15)	End time (text format).
Duration	int	Actual duration of interval in seconds.
IfSpeed	int	Agent interface speed in bits per second.
NetType	char(4)	Agent network type: ET, TR, FDDI, or WAN.
Reinstalled	int	Reinstalled flag, non-zero if reinstalled.
Filler1	int	Reserved for future use.
Utilization	real	Average segment utilization for the time interval.
Octets	real
Pkts	real
Filler2	real	Reserved for future use.
Filler3	real	Reserved for future use.

Specifying Report Parameters

To run reports from the command line using the nsreport command, you must specify the agent, domain, period (start and end time) and report type you want to run. You can also specify optional parameters such as output, database, format, and any additional configuration files or report files that contain specify parameters you want to apply.

You must specify the following parameters:

-agent agent-name [agent-name...] |
@agent-group-name
-switch switch-name | switch-name port-name
[port-name...]
-framerelay framerelay-agentname
framerelay-agentname
dlci-name [dlci-name...]
-domain domain-name [domain-name...]
-period 9701070800 9701080000 //YYMMDDHHMM
-interval00:01 | 00:05 | 00:15 | 00:30 |01:00 |
02:00 | 04:00 | 04:00 | 24:00
-type
segment-summary-et
segment-summary-tr
segment-summary-wan
segment-summary-fddi
segment-history-et
segment-history-tr
segment-history-wan
segment-history-fddi
conversation-summary
conversation-history
host-summary
host-history
vlan-usage
dlci-usage
router-backbone-usage
multi-segment-history-mix
multi-segment-summary-et (Specify any one stat)
multi-segment-summary-tr (Specify any one stat)
multi-segment-summary-wan (Specify any one stat)
multi-segment-summary-fddi (Specify any one stat)
network-availability-summary (%Network Link Down/one segment)
network-availability-history (%Network Link Down/one segment)
network-response-time (Average Round Trip Delay)
wan-usage
protocol-usage (ASCII format only)
protocol-distribution (ASCII format only)
host-verbose (ASCII format only)
host-outbound (ASCII format only)
billing (ASCII format only)
Optional Parameters:
-output[screen | [[print | file | appendfile |
mail] name] ]
-databasedetail | summary
-format[plot | bar | sbar | pie | tabular | 
csv | tsv ]
-cfg[filename | *.rt | *.rd | *.rr ]
-titleoptional-title (Overrides the default report title)
-header1optional-header1
-header2optional-header2
-footeroptional-footer
-toptopN-value
-sortstat-id
-treedomain-tree-file (If other than domtree.inf)
-departmentdepartment-file (Optional)
-tarifftariff-file (Default equals $0.01 per kilobyte)
 
-lpplines-per-page
-statstat-id (Determines which stat is displayed in the first graph. Only on multisegment-history/summary reports.)

Posted: Fri Mar 31 09:12:04 PST 2000
Copyright 1989 - 2000©Cisco Systems Inc.