Job Management Partner 1/Performance Management Planning and Configuration Guide

Appendix C.2 Exclusively occupied disk space

The exclusively occupied disk space sometimes changes according to the number of records that collect the performance data.

The following gives estimates of the exclusively occupied disk space for Performance Management.

Organization of this subsection

(1) Exclusively occupied disk space for the entire system

(2) Exclusively occupied disk space for the Store database (Master Store)

(3) Exclusively occupied disk space for the Store database (health check agent)

(1) Exclusively occupied disk space for the entire system

The exclusively occupied disk space for the entire system is shown below for each OS.

(a) In Windows

Table C-3 Exclusively occupied disk space for the entire system (in Windows)

Program Status Exclusively occupied disk space (unit: megabytes)

Windows Server 2003 Windows Server 2008

PFM - Manager During an installation 155 160

Normal 461 + 0.015 x a x b + 0.01 x c + 2 x d + W 466 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W

PFM - Web Console During an installation 300 140^#+ 355

Normal 300 + 160 + (e x f x g) 140^# + 355 + 160 + (e x f x g)

PFM - Base During an installation 30 35

Normal 112 + 2 x d 117 + 2 x d

Legend:

a: Number of connected PFM - Agents or PRM - RM (if instances are started, the number of instances is also added)

b: Number of saved PA records (initial status is 1,000)

c: Number of user-created reports

d: Upper limit of common message log file

e: Value of logFileNumber in config.xml

f: Value of logFileSize in config.xml

g: Number of commands including jpcrpt and jpcmkkey and subcommands of jpcrdef, jpcasrec, and jpcaspsv (maximum number: 9)

W: Exclusively occupied disk space used for the Store database of the health check agent

#

Exclusively occupied disk space on the system drive. For example, when the system drive is C: and the ordinary drive is D:, 140 MB is used on the C: drive, even if PFM - Web Console is installed on the D: drive.

(b) In UNIX

Table C-4 Exclusively occupied disk space for the entire system (in UNIX)

Program Status Exclusively occupied disk space (unit: megabytes)

HP-UX Solaris
(SPARC) Solaris
(excluding SPARC) AIX Linux
(excluding IPF) Linux
(IPF)

PFM - Manager During an installation -- 220 x 2 -- 410 x 2 220 x 2 --

Normal -- 526 + 0.1 x a+ 0.015 x a x b + 0.01 x c + 2 x d + W -- 716 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W 526 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W --

PFM - Web Console During an installation -- -- -- -- 480 --

Operation in the initial status -- -- -- -- 480 + 160 + e x f x g --

PFM - Base During an installation 240 x 2 50 x 2 50 x 2 220 x 2 70 x 2 80 x 2

Normal 322 + 2 x d 132 + 2 x d 132 + 2 x d 302 + 2 x d 152 + 2 x d 162 + 2 x d

Legend:

a: Number of connected PFM - Agents or PRM - RM (if instances are started, the number of instances is also added)

b: Number of saved PA records (initial status is 1,000)

c: Number of user-created reports

d: Upper limit of common message log file

e: Value of logFileNumber in config.xml

f: Value of logFileSize in config.xml

g: Number of commands including jpcrpt and jpcmkkey and subcommands of jpcrdef, jpcasrec, and jpcaspsv (maximum number: 9)

W: Exclusively occupied disk space used for the Store database of the health check agent

--: Not supported

Program	Status	Exclusively occupied disk space (unit: megabytes)
Windows Server 2003	Windows Server 2008
PFM - Manager	During an installation	155	160
Normal	461 + 0.015 x a x b + 0.01 x c + 2 x d + W	466 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W
PFM - Web Console	During an installation	300	140^#+ 355
Normal	300 + 160 + (e x f x g)	140^# + 355 + 160 + (e x f x g)
PFM - Base	During an installation	30	35
Normal	112 + 2 x d	117 + 2 x d

Program	Status	Exclusively occupied disk space (unit: megabytes)
HP-UX	Solaris (SPARC)	Solaris (excluding SPARC)	AIX	Linux (excluding IPF)	Linux (IPF)
PFM - Manager	During an installation	--	220 x 2	--	410 x 2	220 x 2	--
Normal	--	526 + 0.1 x a+ 0.015 x a x b + 0.01 x c + 2 x d + W	--	716 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W	526 + 0.1 x a + 0.015 x a x b + 0.01 x c + 2 x d + W	--
PFM - Web Console	During an installation	--	--	--	--	480	--
Operation in the initial status	--	--	--	--	480 + 160 + e x f x g	--
PFM - Base	During an installation	240 x 2	50 x 2	50 x 2	220 x 2	70 x 2	80 x 2
Normal	322 + 2 x d	132 + 2 x d	132 + 2 x d	302 + 2 x d	152 + 2 x d	162 + 2 x d

(2) Exclusively occupied disk space for the Store database (Master Store)

The exclusively occupied disk space required for tasks in the Store database of the Master Store service of PFM - Manager is shown below for each OS.

(a) In Windows

Table C-5 Exclusively occupied disk space required for tasks in the Store database (in Windows)

Tasks Estimation expression for exclusively occupied disk space (unit: megabytes)

Backup of Store database 2 x (1 + 0.1 x a + 0.015 x a x b)

Export of Store database 2 x (1 + 0.1 x a + 0.015 x a x b)

Legend:

a: Number of connected PFM - Agents or PRM - RM (if instances are started, the number of instances is also added)

b: Number of saved PA records (initial status is 1,000)

(b) In UNIX

Table C-6 Exclusively occupied disk space required for tasks in the Store database (in UNIX)

Tasks Estimation expression for exclusively occupied disk space (unit: megabytes)

Solaris AIX Linux

Backup of Store database 2 x (1 + 0.1 x a + 0.015 x a x b) 2 x (1 + 0.1 x a + 0.015 x a x b) 2 x (1 + 0.1 x a + 0.015 x a x b)

Export of Store database 2 x (1 + 0.1 x a + 0.015 x a x b) 2 x (1 + 0.1 x a + 0.015 x a x b) 2 x (1+0.1 x a + 0.015 x a x b)

Legend:

a: Number of connected PFM - Agents or PRM - RM (if instances are started, the number of instances is also added)

b: Number of saved PA records (initial status is 1,000)

Tasks	Estimation expression for exclusively occupied disk space (unit: megabytes)
Backup of Store database	2 x (1 + 0.1 x a + 0.015 x a x b)
Export of Store database	2 x (1 + 0.1 x a + 0.015 x a x b)

Tasks	Estimation expression for exclusively occupied disk space (unit: megabytes)
Solaris	AIX	Linux
Backup of Store database	2 x (1 + 0.1 x a + 0.015 x a x b)	2 x (1 + 0.1 x a + 0.015 x a x b)	2 x (1 + 0.1 x a + 0.015 x a x b)
Export of Store database	2 x (1 + 0.1 x a + 0.015 x a x b)	2 x (1 + 0.1 x a + 0.015 x a x b)	2 x (1+0.1 x a + 0.015 x a x b)

(3) Exclusively occupied disk space for the Store database (health check agent)

When using the health check function, you also need to consider the exclusively occupied disk space used by the Store database of the health check agent. The disk space occupied by the Store database depends on the Store version. By default, the health check agent uses Store 2.0. For details on estimating the disk space used by Store 2.0, see C.4 Exclusively occupied disk space when using the Store database in PFM - Agent 08-00 or later (Store 2.0). The exclusively occupied disk space when using Store 1.0 is shown below.

(a) Estimation expression (Store 1.0)

For Store databases, each record is stored by type in a single file. The following table describes the exclusively occupied disk space for Store databases (Store 1.0) by record type.

Table C-7 Exclusively occupied disk space for Store databases by record type

Record type Estimation expression for exclusively occupied disk space (unit: bytes)

PI record type (X₁ + ... + X_a + 3,500 x a)

PD record type (Y₁ + ... + Y_b + 700 x b)

Legend:

X: Exclusively occupied disk space for each type of historical data record for PI-type records

The formula to calculate X is as follows:

X = {e x f + (d + 1,900) x {(e x f) / (65,250 - d) + 1}^#1} x g x 1.5

Y: Exclusively occupied disk space for each type of historical data record for PD-type records

The formula to calculate Y is as follows:

Y = {e x h + (d + 1,900) x {(e x f) / (65,250 - d) + 1}^#1x (h / f)^#2} x 1.5

a: The number of historical data records for a PI-type record

b: The number of historical data records for a PD-type record

d: Size of the fixed part of each type of historical data record^#3

e: Size of the variable part of each type of historical data record^#3

f: Number of instances of each type of historical data record (or 1 if single-instance records)^#5

g: The number of stored records for each type of historical data record (the number of instances is not considered)^#4

h: The maximum number of stored records for each type of historical data record

#1:

Truncate the result of {(e x f) / (65,250 - d) + 1} at the decimal point.

#2:

Truncate the result of (h / f) at the decimal point.

#3:

For details on the sizes of the fixed and variable parts of each record, see H.1 Data model.

#4:

For PI-type records, the collected data is automatically summarized by segment (hour, day, week, month, and year). When calculating the exclusively occupied disk space, you must consider the number of stored records per minute, hour, day, week, month, and year. The following table shows the default retention periods and the number of stored records:

Data Type Retention period Number of stored records
(if the collection interval is 1 minute)

Per-minute 1 day 1,440

Hourly 7 days 168

Daily 1 year 366

Weekly 1 year 52

Monthly 1 year 12

Yearly Indefinite collection-years x 1

#5:

The following table shows the number of instances for each record.

Table C-8 Number of instances for each record

Record ID Command Note

PI_HAVL jpctool service list -id "?A*" -host "*" The number of hosts displayed when this command is entered

PD_HC jpctool service list -id "?A*" -host "*" The number of the Agent Collector and Remote Monitor Collector services displayed when this command is entered

PD_HOST jpctool service list -id "?A*" -host "*" The number of hosts displayed when this command is entered

(b) Example estimate
Example estimate 1

Estimation expression for the PI-type record Host Availability (PI_HAVL), when only PI_HAVL has been set to be saved. The values for variables a-g indicated in Table C-7 are assumed as follows:
a = 1
d = 681
e = 292
f = 25
g = 2,039
(How to determine the value f)

Using Table C-8, you can determine the number of instances for the record. Here the value f is 25.
(How to determine the value g)

Assuming that the collection interval of PI_HAVL is 5 minutes, the yearly collection period is 1 year, and the retention periods are configured as described in Legend #4 for Table C-7.
288 + 168 + 366 + 52 + 12 + 1 = 887 records
g = 887
The formula to calculate X:
X = {e x f + (d + 1,900) x {e x f/(65,250 - d) + 1}} x g x 1.5
X = {292 x 25 + (681 + 1,900) x {292 x 25/(65,250 - 681) +1}} x 887 x 1.5
 = {7,300 + (2,581) x {7,300 / 64,569 + 1}} x 1330.5
 = {7,300 + 2,581 x 1} x 1330.5
 = 9,881 x 1330.5
= 13,146,671 (bytes)
= Approx. 12.5 (megabytes)
Example estimate 2

Estimation expression for the PD-type record Health Check Detail (PD_HC), when only PD_HC has been set to be saved.
b = 1
d = 716
e = 988
f = 25
h = 15,000
(how to determine the value f)

Refer to the number of instances for the performance object that corresponds to the records in Table C-8. Here the value f is 3.

(how to determine the value h)

By default, the maximum number of records for the retention period is 10,000 records.

Here 15, 000 records is set as the maximum number of records for the retention period.
The formula to calculate Y:
Y = {e x h + (d + 1,900) x {(e x f)/(65,250 - d) + 1} x (h/f)} x 1.5 
Y ={988 x 15,000 + (716 + 1,900) x {(988 x 25)/(65,250 - 716) + 1} x (15,000/25)} x 1.5
  = {988 x 15,000 + 2,616 x {24,700 / 64,534 + 1} x 600} x 1.5
  = {14,820,000 + 2,616 x 1 x 744} x 1.5
  = {14,820,000 + 1,202,304} x 1.5
  = 16,022,304 x 1.5
  = 24,033,456 (bytes)
  = Approx. 23 (megabytes)

Record type	Estimation expression for exclusively occupied disk space (unit: bytes)
PI record type	(X₁ + ... + X_a + 3,500 x a)
PD record type	(Y₁ + ... + Y_b + 700 x b)

Data Type	Retention period	Number of stored records (if the collection interval is 1 minute)
Per-minute	1 day	1,440
Hourly	7 days	168
Daily	1 year	366
Weekly	1 year	52
Monthly	1 year	12
Yearly	Indefinite	collection-years x 1

Record ID	Command	Note
PI_HAVL	jpctool service list -id "?A" -host ""	The number of hosts displayed when this command is entered
PD_HC	jpctool service list -id "?A" -host ""	The number of the Agent Collector and Remote Monitor Collector services displayed when this command is entered
PD_HOST	jpctool service list -id "?A" -host ""	The number of hosts displayed when this command is entered

[Contents][Back][Next]

[Trademarks]