Job Management Partner 1/Performance Management - Agent Option for Virtual Machine Description, User's Guide and Reference

[Contents][Glossary][Index][Back][Next]

Appendix A.2 Disk occupancy

Disk occupancy varies according to the number of records from which performance data is collected.

The procedure for estimating the disk occupancy by PFM - Agent for Virtual Machine follows.

Organization of this subsection
(1) Disk occupancy in the entire system
(2) Disk occupancy of the Store database (Store version 1.0)
(3) Disk occupancy of the Store database (Store version 2.0)

(1) Disk occupancy in the entire system

This subsection describes estimated disk occupancy in the entire system.

The disk occupancy by PFM - Agent for Virtual Machine varies according to the file size of the following data:

The table below shows estimates of the disk occupancy in the entire system.

Table A-3 Disk occupancy in the entire system

State of PFM - Agent for Virtual Machine Disk occupancy (megabytes)
Windows Server 2003 / Windows Server 2008
During installation#1 7
After setup#2 8
During operations a + b

Legend:

a:
Disk occupancy during installation

b:
Sum of disk occupancy for all instances
The formula for estimating the disk occupancy of a single instance follows:
25 + c + d

c:
Collected log size
Default: 128 megabytes (16 megabytes x 8 sectors)
Maximum value: 256 megabytes (32 megabytes x 8 sectors)

d:
Disk occupancy of the Store database
For details about the disk occupancy of the Store database, see (2) Disk occupancy of the Store database (Store version 1.0) or (3) Disk occupancy of the Store database (Store version 2.0).

#1
During installation, disk space that is twice the size of the main program is required.

#2
Indicates a situation in which only one instance of PFM - Agent for Virtual Machine, for which no collection is specified, has been set up.

(2) Disk occupancy of the Store database (Store version 1.0)

This subsection explains the disk occupancy of the Store database (Store version 1.0).

(a) Estimation formulas

In the Store database, a record of each type is stored in a file. Table A-6 shows the disk occupancy of the Store database (Store version 1.0) for each record type.

Notes:

1. When performance data is stored in the Store database, several fields are added. Because the additional fields are included in the disk occupancy, there is no need to estimate the size anew.
  • Fields that are added to all records
    The table below shows the fields that are added to all records.

    Table A-4 Fields that are added to all records

    PFM - View name PFM - Manager
    name    		 Explanation
    Agent Host DEVICEID Name of the host on which PFM - Agent is running
    Agent Instance PROD_INST PFM - Agent instance name
    Agent Type PRODID PFM - Agent product ID
    Date DATE Date on which the record was created (Greenwich Mean Time)
    Date and Time DATETIME Combination of the Date (DATE) and Time (TIME) fields
    Drawer Type DRAWER_TYPE Data summarization category (minute, hour, day, week, month, or year) for a record of the PI record type
    GMT Offset GMT_ADJUST Difference between Greenwich Mean Time and the local time (in seconds)
    Time TIME Time at which the record was created (Greenwich Mean Time)
  • Fields that are added when data of the PI record type is summarized
    The following fields, shown below as character strings, are added as suffixes to the PFM - View name and the PFM - Manager name. The table below shows the fields that are added when data of the PI record type is summarized.

    Table A-5 Fields that are added when data of the PI record type is summarized

    PFM - View name PFM - Manager name Explanation
    PFM - View-name(Total) PFM - Manager-name_TOTAL Field value total
    -- PFM - Manager-name_COUNT Collected record count
    PFM - View-name(Max) PFM - Manager-name_HI Maximum field value
    PFM - View-name(Min) PFM - Manager-name_LO Minimum field value

    Legend:
    --: Not applicable

  • Fields that are added when data stored in the Store database is exported using the jpctool db dump (jpcctrl dump) command
    When data stored in the Store database is exported using the jpctool db dump (jpcctrl dump) command, the fields listed below are output. These fields are also added when data is stored in the Store database. Because these fields are used internally by PFM - Agent for Virtual Machine, do not use them for regular operations.
    [Figure] record-ID_DATE_F
    [Figure] record-ID_DEVICEID_F
    [Figure] record-ID_DRAWER_TYPE_F
    [Figure] record-ID_DRAWER_COUNT
    [Figure] record-ID_DRAWER_COUNT_F
    [Figure] record-ID_INST_SEQ
    [Figure] record-ID_PFM - Manager-name_COUNT
    [Figure] record-ID_PFM - Manage-name_SEC
    [Figure] record-ID_PFM - Manager-name_MSEC
    [Figure] record-ID_PRODID_F
    [Figure] record-ID_PROD_INST_F
    [Figure] record-ID_RECORD_TYPE
    [Figure] record-ID_RECORD_TYPE_F
    [Figure] record-ID_SEVERITY
    [Figure] record-ID_SEVERITY_F
    [Figure] record-ID_TIME_F
    [Figure] record-ID_UOWID
    [Figure] record-ID_UOWID_F
    [Figure] record-ID_UOW_INST
    [Figure] record-ID_UOW_INST_F

2. When you execute the jpctool db backup (jpcctrl backup) or jpctool db dump (jpcctrl dump) command, the backup file or export file requires approximately twice the disk space as that calculated using the formula in the table below.

Table A-6 Disk occupancy of the Store database by record type

Record type Disk occupancy estimation formula (bytes)
PI record type X1 + ... + Xa + 3,500 x a
PD record type Y1 + ... + Yb + 700 x b

Legend:
X: Disk occupancy of each record that collects historical data as PI record type records
The calculation formula for X follows.
X = {d x e + (c + 1,900) x {(d x e)/(65,250 - c) + 1}#1} x f x 1.5
Y: Disk occupancy of each record that collects historical data as PD record type records
The calculation formula for Y follows.
Y = {d x g + (c + 1,900) x {(d x e)/(65,250 - c) + 1}#1 x (g/e)#2} x 1.5
a: Number of records that collect historical data as PI record type records
b: Number of records that collect historical data as PD record type records
c: Size of the fixed portion of each record that collects historical data#3
d: Size of the variable portion of each record that collects historical data#3
e: Number of instances in each record that collect historical data#7
(This value is 1 for single-instance records)
f: Number of records saved for each record that collects historical data (the instance count is not taken into consideration)#4, #5
g: Upper limit for the number of records saved for each record that collects historical data#6

#1
For the calculation result of {(d x e)/(65,250 - c) + 1}, discard the decimals.

#2
For the calculation result of (g/e), discard the decimals.

#3
For details about the size of the fixed and variable portions of each record, see 5. Records.

#4
In the case of PI record type records, since the collected data is automatically summarized according to a certain category (by hour, day, week, month, and year), you must take into consideration the number of records saved for the minute, hour, day, week, month, and year categories when you calculate the number. The table below shows the default retention period and the number of records to be saved.

Table A-7 Data types and retention period of PI record type and the number of records to be saved

Record type Data type Retention period Number of records to be saved (when the collection interval is 1 minute)
PI Every minute 1 day 1,440
Every hour 7 days 168
Every day 1 year 366
Every week 1 year 52
Every month 1 year 12
Every year No limit (number of collection years) x 1

#5
For PD record type records, specify the number of records to be saved. The table below shows the default number of records that are saved.

Table A-8 Number of PD record type records saved

Record type Number of records saved
PD 10,000

#6
For the number of records to be saved, see E.1 Agent Store service property list.

#7
The table below shows how to calculate the instance count for each record.

Table A-9 Instance count for each record

Record ID Virtual environment Instance count calculation method Verification method
PD VMware 1 (because this is a single-instance record) --
Hyper-V --
PD_VM VMware Number of virtual machines on the connected physical server #8
Hyper-V #9
PI VMware 1 (because this is a single-instance record) --
Hyper-V --
PI_VI VMware Number of virtual machines running on the connected physical server #8
Hyper-V #9
PI_HCI VMware Number of physical CPUs on a physical server #8
Hyper-V #9
PI_VCI VMware Total number of CPUs allocated to the virtual machines running on the connected physical server #8
Hyper-V #9
PI_HMI VMware 1 (because this is a single-instance record) --
Hyper-V --
PI_VMI VMware Number of virtual machines running on the connected physical server #8
Hyper-V #9
PI_HPDI VMware Total number of physical disks on a physical server #8
Hyper-V #9
PI_VPDI VMware Total number of physical disks allocated to the virtual machines running on the connected physical server #8
Hyper-V #9
PI_HLDI VMware Total number of logical disks on a physical server #8
Hyper-V #9
PI_VLDI VMware Total number of logical disks secured by the virtual machines running on the connected physical server #10
Hyper-V Not supported --
PI_HNI VMware Total number of physical network cards on a physical server #8
Hyper-V #9
PI_VNI VMware Total number of network cards of the virtual machines running on the connected physical server #8
Hyper-V #9

Legend:
--: Not applicable

#8
You can use Virtual Infrastructure Client to verify the instance count for the record; Table A-10 shows how to do this. For details about how to use Virtual Infrastructure Client, see the Help section of Virtual Infrastructure Client.

#9
You can use Windows Hyper-V Manager and the System Monitor snap-in of the Microsoft Management Console to verify the instance count for the record; Table A-10 shows how to do this.

#10
You can use Managed Object Browser to verify the instance count for the PI_VLDI record. Note that in order to check the instance count, the virtual machine must be running.
Table A-10 shows how to use Managed Object Browser to verify the instance count. For details about how to use Managed Object Browser, see the VMware documentation.

Table A-10 Instance count verification method

Record ID Virtual environment Calculation method
PD_VM VMware From the Virtual Infrastructure Client tree, select a host and then choose the Virtual Machine tab.
The instance count is the displayed number of virtual machines.
Hyper-V At the Hyper-V Manager, select a Hyper-V server and check the number of virtual machines displayed in the Virtual machines pane.
The instance count is the displayed number of virtual machines plus 1 (root partition).
PI_VI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Virtual Machine tab.
The instance count is the number of virtual machines whose State is Power On.
Hyper-V At the Hyper-V Manager, select a Hyper-V server and check the number of virtual machines displayed in the Virtual machines pane.
The instance count is the number of virtual machines whose State is Running plus 1 (root partition).
PI_HCI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Summary tab.
The instance count is the value displayed for Processors under General.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Processor object.
The instance count is the number of instances displayed in Instances of the selected object minus _Total minus all-instances.
PI_VCI VMware From the Virtual Infrastructure Client tree, select a virtual machine and then choose the Summary tab.
The instance count is the value displayed for CPU under General.
Hyper-V The instance count is the number of instances displayed for Processor plus the number of instances displayed for Hyper-V Hypervisor Virtual Processor as described below.

Number of instances for Processor
In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Processor object.
The instance count is the number of instances displayed in Instances of the selected object minus _Total minus all-instances.

Number of instances for Hyper-V Hypervisor Virtual Processor
In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Hyper-V Hypervisor Virtual Processor object.
The instance count is the number of instances displayed in Instances of the selected object minus _Total minus all-instances.
PI_VMI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Virtual Machine tab.
The instance count is the number of virtual machines whose State is Power On.
Hyper-V At the Hyper-V Manager, select a Hyper-V server and check the number of virtual machines displayed in the Virtual machines pane.
The instance count is the number of virtual machines whose State is Running.
PI_HPDI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Performance tab.
From Switch to, select Disk.
The instance count is the number of objects, excluding host names, displayed under Performance Chart Legend.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the PhysicalDisk object.
The instance count is the number of instances displayed in Instances of the selected object minus _Total minus all-instances.
PI_VPDI VMware From the Virtual Infrastructure Client tree, select a virtual machine and then choose the Performance tab.
From Switch to, select Disk.
The instance count is the number of objects, excluding virtual machine names, displayed under Performance Chart Legend.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Hyper-V Virtual Storage Device object.
The instance count is the displayed number of instances.
PI_HLDI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Summary tab.
The instance count is the value displayed for Datastore under Resources.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the LogicalDisk object.
The instance count is the number of instances displayed in Instances of the selected object minus _Total minus all-instances.
PI_VLDI VMware The following describes how to use Managed Object Browser to check the number of instances:
  1. With a Web browser, access the following URL:
    https://monitored-VMware-host-name/mob
  2. In the Methods table, click the RetrieveServiceContent link that is displayed on the ServiceContent row in the NAME column.
  3. Click the Invoke Method link displayed at the lower right of the screen.
  4. In the Method Invocation Result: ServiceContent table, click the ha-folder-root link that is displayed on the rootFolder row in the VALUE column.
  5. In the Properties table, click the ha-datacenter link that is displayed on the childEntity row in the VALUE column.
  6. In the Properties table, click the ha-folder-vm link that is displayed on the vmFolder row in the VALUE column.
  7. In the Properties table, of the values displayed in the VALUE column on the childEntity row, click the link for the virtual machine ID whose number of logical disks is to be checked.
    You can check the virtual machine name in the VALUE column on the name row in the window that is displayed after this operation.
  8. In the Properties table, click the guest link that is displayed on the guest row in the VALUE column.
  9. In the Properties table, click the disk link that is displayed on the disk row in the VALUE column.
  10. Check the displayed number of disks.
Hyper-V Not supported
PI_HNI VMware From the Virtual Infrastructure Client tree, select a host and then choose the Performance tab.
From Switch to, select Network.
The instance count is the number of objects, excluding host names, displayed under Performance Chart Legend.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Network Interface object.
The instance count is the displayed number of instances.
PI_VNI VMware From the Virtual Infrastructure Client tree, select a virtual machine and then choose the Performance tab.
From Switch to, select Network.
The instance count is the number of objects, excluding virtual machine names, displayed under Performance Chart Legend.
Hyper-V In the root partition, start System Monitor or Performance Logs and Alerts of Microsoft Management Console, and then from the Add Counters dialog box, select the Hyper-V Virtual Switch Port object.
The instance count is the displayed maximum number of instances.

(b) Estimation examples

Estimation example 1
The explanation here uses an example in which PI_VI records are collected. It is assumed that variables a-f in Table A-6 are set to the following values:
a = 1
c = 681
d = 857
e = 2
f = 2,039

(How to determine e)
See the instance count of the performance object that corresponds to the record in Table A-9. For this example, e = 2 is selected.

(How to determine f)
When the PI_VI collection interval is 1 minute, the number of collection years with annual collection is 1 year, and the retention settings are as shown in Table A-7:
1,440 + 168 + 366 + 52 + 12 + 1 = 2,039 records
f = 2,039

Calculation formula for X:
X = {d x e + (c + 1,900) x {d x e/(65,250 - c) + 1}} x f x 1.5
X = {857 x 2 + (681 + 1,900) x {857 x 2/(65,250 - 681) + 1}} x 2,039 x 1.5
  = {1,714 + (2,581) x {1,714/64,569 + 1}} x 3,058
  = {1,714 + 2,581 x 1} x 3,058
  = 4,295 x 3,058
  = 13,134,110 (bytes)
  = Approximately 13 megabytes

Estimation example 2
The explanation here uses an example in which PD_VM records are collected.
b = 1
c = 681
d = 369
e = 3
g = 2,232

(How to determine e)
See the instance count of the performance object that corresponds to the record in Table A-9. For this example, e = 3 is selected.

(How to determine g)
The default retention setting is 10,000 records.
When the collection interval is set to 3,600 seconds and one month's worth of data is to be saved:
g = 24 records (one day's worth) x 31 (1 month) x 3 (value of e) = 2,232 records
g = 2,232 records

Calculation formula for Y:
Y = {d x g + (c + 1,900) x {(d x e)/(65,250 - c) + 1} x (g/e)} x 1.5
Y = {369 x 2,232 + (681 + 1,900) x {(369 x 3)/(65,250 - 681) + 1} x (2,232/3)} x 1.5
  = {823,608 + 2,581 x {1,107/64,569 + 1} x 744} x 1.5
  = {823,608 + 2,581 x 1 x 744} x 1.5
  = {823,608 + 1,920,264} x 1.5
  = 2,743,872 x 1.5
  = 4,678,272 (bytes)
  = Approximately 4 megabytes

(3) Disk occupancy of the Store database (Store version 2.0)

This subsection explains the disk occupancy of the Store database (Store version 2.0).

(a) Estimation formulas

This subsection explains how to estimate the disk occupancy, file count, folder count, and the number of files that are opened by the Store service.

n Disk occupancy

The disk occupancy of the Store database is the sum total of the disk occupancy of each record type. For the PI record type, the disk occupancy of the Store database is the sum total of the disk occupancy of the individual summarization categories.

Estimation formula for disk occupancy X for each record type (bytes)
X = {(e + 2) x f' + (d + 60) x {((e + 2) x f')/(65,250 - d) + 1}#1} x a/b x (c + 1) x 1.1
a: The value differs depending on the record type and the summarization category. See Table A-11.
b: The value differs depending on the record type and the summarization category. See Table A-11.
c: Retention period setting for historical data#3. The unit to be specified differs depending on the record type and the summarization category. For the unit to be specified, see Table A-11.
d: Size of the fixed portion of each record that collects historical data#4
e: Size of the variable portion of each record that collects historical data#4
f: Number of instances in each record that collects historical data (this value is 1 for single-instance records)#5
f': If f is 2 or greater, it is rounded up to the nearest multiple of 4. For example, when f = 2, f' = 4. When f = 13, f' = 16. However, when f = 1, f' = 1.

Table A-11 Values to be set for a, b, and c

Record type Summarization category a b c
PI Minute 1,440 1 + (g - 1)/60#2 Retention period (days)
Hour 24 1 + (g - 1)/3,600#2 Retention period (days)
Day 7 1 + (g - 1)/86,400#2 Retention period (weeks)
Week 1 1 + (g - 1)/604,800#2 Retention period (weeks)
Month 1 1 + (g - 1)/2,592,000#2 Retention period (months)
Year 1 1 + (g - 1)/31,622,400#2 Retention period (years)
PD -- 1,440 g/60 Retention period (days)

Legend:
g: Historical data collection interval setting (seconds)
--: Not applicable

#1
For the calculation result of {((e + 2) x f')/(65,250 - d) + 1}, discard the decimals.

#2
For the calculation result of b for the PI record type, discard the decimals.

#3
The table below shows the default retention periods for Store version 2.0.

Table A-12 Default retention period (Store version 2.0)

Record type Data type Retention period
PI Every minute 1 day
Every hour 7 days
Every day 54 weeks
Every week 54 weeks
Every month 1 year
Every year No limit
PD -- 7 days

Legend:
--: Not applicable

#4
For details about the size of the fixed and variable portions of each record, see the section on record size in 5. Records.

#5
For the instance count for each record, see (2) Disk occupancy of the Store database (Store version 1.0).

n File count

The formula for estimating the number of files N created in the Store database follows.

N = 20 + 2 x (
  (A11 + A12 + ... + A1l + l) + 
  (A21 + A22 + ... + A2l + l) + 
  (A31 + A32 + ... + A3l + l) + 
  (A41 + A42 + ... + A4l + l) + 
  (A51 + A52 + ... + A5l + l) + 
  (11 x l) + 
  (B1 + B2 + ... + Bm + m)
  )

l: Number of records collected in the PI record type

m: Number of records collected in the PD record type

A11-A1l: Retention period setting for minute-by-minute records of the PI record type (in days)

A21-A2l: Retention period setting for hourly records of the PI record type (in days)

A31-A3l: Retention period setting for daily records of the PI record type (in weeks)

A41-A4l: Retention period setting for weekly records of the PI record type (in weeks)

A51-A5l: Retention period setting for monthly records of the PI record type (in months)

B1-Bm: Retention period setting for each PD record type record (in days)

n Folder count

The formula for estimating the number of folders N created in the Store database follows.

N = 25 + 2 x ((A1max) + (A2max) + (A3max) + (A4max) + (A5max) + 11 + (Bmax))

A1max: Maximum retention period for PI record type records whose data record summarization category is Minutes (in days)

A2max: Maximum retention period for PI record type records whose data record summarization category is Hours (in days)

A3max: Maximum retention period for PI record type records whose data record summarization category is Days (in weeks)

A4max: Maximum retention period for PI record type records whose data record summarization category is Weeks (in weeks)

A5max: Maximum retention period for PI record type records whose data record summarization category is Months (in months)

Bmax: Maximum retention period for PD record type records (in days)

n Number of files opened by the Store service

The formula for estimating the number of files N opened by the Store service follows.

N = 20 + 2 x (6 x l + m)

l: Number of records collected of the PI record type

m: Number of records collected of the PD record type

(b) Estimation examples

This subsection provides estimation examples for Store database (Store version 2.0) of PFM - Agent for Virtual Machine.

n Disk occupancy

This explanation uses an example in which PI_VI and PD_VM records are collected.

A disk occupancy estimate for PI_VI records follows. The following values of a to g are examined using the disk occupancy estimation formulas in (a) Estimation formulas.

d = 681 bytes
e = 857 bytes
f = 3 is used in this example.
f' = 4 is used in this example.
g = 60 seconds is used in this example.

Next, calculate minute-by-minute records, hourly records, and so on.

Minute-by-minute records
The variables are set to the following values:
a = 1,440
b = 1 + (60 - 1)/60 = 1.98... = 1 (with decimals discarded)
c = 3 days is used in this example.
The estimation formula follows.
X (minutes) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250-681) + 1}} x 1,440/1 x (3 + 1) x 1.1
        = {3,436 + 741 x 1} x 6,336
        = 4,177 x 6,336
        = 26,465,472 (bytes) = Approximately 25 megabytes

Hourly records
The variables are set to the following values:
a = 24
b = 1 + (60 - 1)/3,600 = 1.01... = 1 (with decimals discarded)
c = 3 days is used in this example.
The estimation formula follows.
X (hours) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250-681) + 1}} x 24/1 x (3 + 1) x 1.1
       = {3,436 + 741 x 1} x 105.6
       = 4,177 x 105.6
       = 441,091 (bytes) = Approximately 0.4 megabytes

Daily records
The variables are set to the following values:
a = 7
b = 1 + (60 - 1)/86,400 = 1.00... = 1 (with decimals discarded)
c = 1 week is used in this example.
The estimation formula follows.
X (days) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250-681) + 1}} x 7/1 x (1 + 1) x 1.1
       = {3,436 + 741 x 1} x 15.4
       = 4,177 x 15.4
       = 64,325 (bytes) = Approximately 0.06 megabytes

Weekly records
The variables are set to the following values:
a = 1
b = 1 + (60 - 1)/604,800 = 1.00... = 1(with decimals discarded)
c = 1 week is used in this example.
The estimation formula follows.
X (weeks) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250-681) + 1}} x 1/1 x (1 + 1) x 1.1
       = {3,436 + 741 x 1} x 2.2
       = 4,129 x 2.2
       = 9,189 (bytes) = Approximately 0.01 megabytes

Monthly records
The variables are set to the following values:
a = 1
b = 1 + (60 - 1)/2,592,000 = 1.00... = 1 (with decimals discarded)
c = 1 month is used in this example.
The estimation formula follows.
X (months) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250 - 681) + 1}} x 1/1 x (1 + 1) x 1.1
       = {3,436 + 741 x 1} x 2.2
       = 4,177 x 2.2
       = 9,189 (bytes) = Approximately 0.01 megabytes

Annual records
The variables are set to the following values:
a = 1
b = 1 + (60 - 1)/31,622,400 = 1.00... = 1 (with decimals discarded)
c = 10 years is used in this example.
The estimation formula follows.
X (years) = {(857 + 2) x 4 + (681 + 60) x {((857 + 2) x 4)/(65,250 - 681) + 1}} x 1/1 x (10 + 1) x 1.1
       = {3,436 + 741 x 1} x 12.1
       = 4,177 x 12.1
       = 50,541 (bytes) = Approximately 0.05 megabytes

Based on the above, disk occupancy for PI_VI is estimated as follows.

X (total) = X (minutes) + X (hours) + X (days) + X (weeks) + X (months) + X (years)
          = 25.53 (megabytes)
          = Approximately 26 megabytes

Next, a disk occupancy estimate for PD_VM records follows.

The variables are set to the following values:

a = 1,400
b = 60/60 = 1
c = 7 days is used in this example.
d = 681 bytes
e = 369 bytes
f = 10 is used in this example.
f' = 12 is used in this example.
g = 60 seconds is used in this example.

The estimation formula follows:

X = {(369 + 2) x 12 + (681 + 60) x {((369 + 2) x 12)/(65,250 - 681) + 1}} x 1,440/1 x (7 + 1) x 1.1
  = {4,452 + 741 x 1} x 12,672
  = 5,193 x 12,672
  = 65,805,696 (bytes) = Approximately 63 megabytes

Therefore, the required disk occupancy for PI_VI + PD_VM = 89 megabytes.

n File count

This explanation uses an example in which PI_VI, PD, and PD_VM records are collected. The values for the variables in (a) Estimation formulas are examined.

l: Number of records collected in the PI record type = 1
m: Number of records collected in the PD record type = 2
A11-A1l: Retention period setting for minute-by-minute records of the PI record type (in days) = 3 days in this example.
A21-A2l: Retention period setting for hourly records of the PI record type (in days) = 3 days in this example.
A31-A3l: Retention period setting for daily records of the PI record type (in weeks) = 1 week in this example.
A41-A4l: Retention period setting for weekly records of the PI record type (in weeks) = 1 week in this example.
A51-A5l: Retention period setting for monthly records of the PI record type (in months) = 1 month in this example.
B1-Bm: Retention period setting for each PD record type record (in days) = 10 days in this example.

The formula for estimating the number of files N created in the Store database follows.

N = 20 + 2 x (
  (A11 + A12 + ... + A1l + l) + 
  (A21 + A22 + ... + A2l + l) + 
  (A31 + A32 + ... + A3l + l) + 
  (A41 + A42 + ... + A4l + l) + 
  (A51 + A52 + ... + A5l + l) + 
  (11 x l) +
(B1 + B2 + ... + Bm + m) + 
)
 = 20 + 2 x {
[3 (for PI_VI) + 1] + 
[3 (for PI_VI) + 1] + 
[1 (for PI_VI) + 1] + 
[1 (for PI_VI) + 1] + 
[1 (for PI_VI) + 1] + 
[11 x 1] + 
[10 (for PD) + 10 (for PD_VM) + 2]
}
 = 20 + 2 x {4 + 4 + 2 + 2 + 2 + 11 + 22} = 114

n Folder count

This explanation uses an example in which PI_VI, PD, and PD_VM records are collected.

l: Number of records collected in the PI record type = 1
m: Number of records collected in the PD record type = 2
A1max: Maximum retention period for PI record type records whose data record summarization category is Minutes (in days) = 3 days in this example.
A2max: Maximum retention period for PI record type records whose data record summarization category is Hours (in days) = 3 days in this example.
A3max: Maximum retention period for PI record type records whose data record summarization category is Days (in weeks) = 1 week in this example.
A4max: Maximum retention period for PI record type records whose data record summarization category is Weeks (in weeks) = 1 week in this example.
A5max: Maximum retention period for PI record type records whose data record summarization category is Months (in months) = 1 month in this example.
Bmax: Maximum retention period for PD record type records (in days) = 10 days in this example.
     (Approach used: If the retention period is 8 days for PD_VM and 10 days for PD, 10 days is used.)

The formula for estimating the number of files N opened by the Store service follows.

N = 25 + 2 x ((A1max) + (A2max) + (A3max) + (A4max) + (A5max) + 11 + (Bmax))
  = 25 + 2 x (3 + 3 + 1 + 1 + 1 + 11 + 10) = 85

n Number of files opened by the Store service

This explanation uses an example in which PI, PD, and PD_VM records are collected.

l: Number of records collected of the PI record type = 1
m: Number of records collected of the PD record type = 2

The formula for estimating the number of files N opened by the Store service follows.

N = 20 + 2 x (6 x l + m)
  = 20 + 2 x (6 x 1 + 2) = 36

[Contents][Back][Next]

[Trademarks]

All Rights Reserved. Copyright (C) 2009, 2010, Hitachi, Ltd.