System Health provides a system wide performance view of SAN resources in the server group.
System Health provides:
o Status of managed resources requiring attention
o Bandwidth measurements of ports, disk pools and physical disks
o Storage capacity breakdowns
o Performance Spotlight which shows key performance statistics for virtual disks
To open System Health:
In the Ribbon>Home tab, click System Health in the Diagnostics area.
Also see:
The Status tab in System Health tool provides a quick visual summary of the overall health of resources managed by SANsymphony software. Internal system monitors report events when status changes for a managed resource. Events occurring in the SAN are classified by severity, time-stamped and reported in the tool. Status of a resource is presented in a single message and is displayed according to the severity; higher severities are displayed before lesser severities. At a glance, any managed resource requiring attention can be identified in the tool. Double-click on a message in the list to view details of the object. This tool can be kept open and active in your workspace to continually monitor the health of your SAN resources.
Physical disks with S.M.A.R.T. (also known as SMART) implementation are monitored for predicted drive failure.
Security Classifications (highest to lowest):
|
Critical. This resource has received a status which is unhealthy and requires immediate attention. |
|
Warning. This resource has received a status which requires attention. Check this resource and take action to ensure it does not reach a critical status. |
|
Attention. This resource may require attention. Verify the status condition. |
Each message may contain multiple reports containing different severity classifications; the message will be displayed under the highest severity classification. For instance, If a Fibre Channel port is Not-connected (Attention severity) and the Loop/Link is down (Critical severity), then this message is reported under the Critical classification at the top of the list.
System Health messages are displayed as long as they are valid. If the System Health tool is open and no messages are displayed, then all managed resources are healthy.
Link errors have to
be manually reset in order to remove them from System Health. See Resetting
Link Error Counts.
The Bandwidth tab in System Health provides measurements associated with bandwidth to provide an overall view of network traffic. This tab can be used to identify specific SAN resources using excessive bandwidth. Measurements are provided for all front-end and mirror ports, disk pools, and physical disks for each DataCore Server in the server group. Data is refreshed frequently. Throughput, latency, and I/Os per second for each SAN resource is provided in real time.
Expand the objects in the tree to view all measurements for the DataCore Server. Comparatively high latency figures for a SAN resource may indicate potential network issue and identify where to begin troubleshooting in the server group when a bottleneck is suspected.
Bandwidth Measurements:
o Throughput
· Front-end Throughput = Amount of data received by the target port.
· Mirror Throughput = Amount of data received by the mirror target port. (Data sent by a port as an initiator is not counted.)
· Back-end Throughput = Amount of data sent by the initiator port.
· Physical Disk Throughput = Time of I/O completion. This is on the Initiator end and includes the transfer layer time.
· Disk Pool Throughput = Time of I/O completion.
o Latency is the largest delay in data transmission and displayed in milliseconds.
o IO/s (IOs per second) is the average number of read and write operations performed.
The Storage tab in System Health provides pie charts for a quick visual representation of the storage in the server group. The Licensed Storage Capacity pie chart represents the regular (non-bulk) storage in the group. The Licensed Bulk Storage Capacity pie chart is displayed when there is bulk storage licensed in the group. See Server Group Details for server group licensing.
Licensed Storage Capacity is the amount of storage that is licensed for the server group. Licensed storage capacity includes allocated, free, and unconfigured storage. DataCore Disks are not included in the Licensed Storage Capacity view.
Total Storage Capacity is the amount of storage that is licensed for the server group including DataCore Disks. Total storage capacity includes allocated, free, unconfigured, and DataCore Disks (allocated and free). To view the total storage capacity, select the Show DataCore Disks check box in the Storage tab.
The following storage categories can be displayed in the Licensed Storage Capacity pie chart:
The different storage
categories displayed in the pie chart are calculated and displayed as
a percentage of the entire amount. Percentages are rounded to the nearest
whole percent. A category must be at least 1% of the whole in order to
be displayed in the pie chart. The different categories of storage
are color-coded in a key next to the chart.
o Allocated storage is the amount of physical storage that is currently in use. This includes storage allocation units (SAUs) that have been allocated from disk pools and used by pass-through disks that are being managed by the software.
o DataCore Disks - Free is the amount of storage in DataCore Disks that is not allocated.*
o DataCore Disks - Allocated
is the amount of storage in DataCore Disks
that is currently in use as storage.*
*DataCore Disks
are virtual disks created by DataCore Servers
and served to DataCore Servers
in the same server group where they may be used as storage. This category
will only be displayed in the Total storage
capacity view of the chart when the Show
DataCore Disks
check box has been selected.
o Expansion limit is the percentage in use (allocated and free) of the maximum amount of licensed storage for the server group at the current license level.
o Free storage is the amount of storage that has been added to pools and is available for use, but not allocated.
o Unconfigured storage is the difference between the amount of licensed storage and the amount of storage that is currently configured (allocated and free storage) in disk pools and pass-through disks. For instance, if the licensed storage capacity is 10 TB and the amount of configured storage (allocated and free) is 2 TB, the unconfigured storage would be 8 TB.
The following storage categories can be displayed in the Licensed Bulk Storage Capacity pie chart:
The different storage
categories displayed in the pie chart are calculated and displayed as
a percentage of the entire amount. Percentages are rounded to the nearest
whole percent. A category must be at least 1% of the whole in order to
be displayed in the pie chart. The different categories of storage are
color-coded in a key next to the chart. (Bulk storage has no expansion
limit.)
o Allocated storage is the amount of physical bulk storage that is currently in use. This includes storage allocation units (SAUs) that have been allocated from bulk disk pools that are being managed by the software.
o Free storage is the amount of storage that has been added to bulk disk pools and is available for use, but not allocated.
o Unconfigured storage is the difference between the amount of licensed storage and the amount of storage that is currently configured (allocated and free storage) in disk pools. For instance, if the licensed storage capacity is 10 TB and the amount of configured storage (allocated and free) is 2 TB, the unconfigured storage would be 8 TB.
The Performance Spotlight feature displays key performance statistics for the top 20 virtual disks in the server group experiencing the highest workload. This tool enables an administrator to quickly view the virtual disks with the highest workload in order to identify potential problems, investigate the cause, and intervene before performance degradation is noticeable.
A sampling of performance counter data is collected by the software per virtual disk in two minute intervals over time so that averages of the historical data can be calculated for a selected time period. Counter data is calculated and presented in three performance statistics intended to alert administrators to possible traffic bottlenecks in the configuration: Average Latency (ms), Average Bytes Transferred/sec, and Average Operations/sec.
The Performance Spotlight statistics are presented in bar chart and table views. Data displayed in the tool can be retrieved based on any one of the statistics and can be sorted by any one of these statistics. For example, data can be retrieved for the top 20 virtual disks with the highest average latency, but be sorted by the virtual disks that have the highest average number of bytes transferred per second.
Data can be displayed for the last seven, 30, or 90 days in order to see the values for the same data in a progression over time. Data for a selected time period can also be exported to a CSV (comma-separated value) file for use with other applications.
These statistics are derived by DataCore performance counters:
o Average Latency in milliseconds
- the sum of the time spent servicing read and write requests during the
selected time period for a given virtual disk, divided by the total number
of read and write requests made during that time period.
The calculation is based on the performance of all the individual paths
of the virtual disk. The time spent servicing read and write requests
is calculated by adding all values for the TotalReadTime and TotalWriteTime
counters for each path* during the selected time period. The total number
of read and write requests is calculated by adding all values for TotalReads
and TotalWrites counters during the selected time period.
The value displayed per host in the table view shows the average latency
for a given path between a host and a virtual disk.
o Average
Bytes Transferred per second - the sum of all read and write data
bytes during the selected time period for a given virtual disk, divided
by the number of seconds in that time period.
The calculation is based on the performance of all the individual paths
of the virtual disk. The sum of the data bytes is calculated by adding
all values for the TotalBytesRead and TotalBytesWritten counters for each
path* during the selected time period.
o Average
Operations per second - the sum of all read and write requests
during the selected time period for a given virtual disk, divided by the
number of seconds in that time period.
The calculation is based on the performance of all the individual paths
of the virtual disk. The sum of the read and write requests is calculated
by adding all values for the TotalReads and TotalWrites counters for each
path* during the selected time period.
*A path is a mapping between an initiator and target port of a logical disk (virtual disk storage source) on a server, In Live Performance, the paths are defined as virtual logical units and are identified by the initiator port, target port, and LUN.
Important Notes:
o Counter data is sampled at two minute intervals. Calculations are based on the sampled data over the course of the selected time period.
o The statistics in the tool are based on the performance metrics of the individual paths of the virtual disk and only take into consideration virtual disks that are currently served to hosts. Snapshots, rollbacks, and virtual disks served via loopback ports and to other DataCore Servers are also taken into consideration.
o The top 20 virtual disks that appear in the tool are chosen based on the selected statistic (Average Latency (ms), Average Bytes Transferred/sec or Average Operations/sec).
o When data is incomplete for a virtual disk in the selected time period, the actual data for that time period will be displayed. Data could be incomplete if a virtual disk has been served to a host for less than the selected time period, or if paths have been in existence for less than the selected time period. For example, if the selected time period is Last 7 days and a virtual disk was served to a host for two days of that time period, the statistic will be calculated using only the data that is available, but averaged over the entire time period of seven days. The names of virtual disks with incomplete data for the selected time period are displayed in italics. Names and data can be displayed in a tool tip when the virtual disk is pointed to by the mouse.
o Virtual disks with no performance data will not have values displayed. This is most like likely to occur when a virtual disk has just been served to a host and performance data has not yet been recorded.
o Data in the tool is automatically updated every six hours, to manually refresh data, click Refresh in the System Health Actions tab in the Ribbon.
o Historical recording, an automatic recording of key performance data over time, is required for this tool to display data. If tool selections are unavailable and data is not displayed in the tool, click the link at the top of the tab to re-enable historical recording. The link will only be displayed if historical recording is disabled. See Historical Performance Monitoring.
The chart view is a graphical presentation of the three Performance Spotlight statistics in individual bar charts.
The statistic that is used to select the top 20 virtual disks is selected in the chart view. The set of virtual disks in all charts are the same, selected by the statistic. The top chart displayed in the tool is the chart for the selected statistic. Each bar in the chart represents a virtual disk. Charts for the two statistics not selected are displayed under the top chart. The bottom charts are in alignment with the top chart so that the bars representing the same virtual disks are always in the same alignment. In this manner, all statistics for the same virtual disk can be viewed in the tab by scrolling up and down. By default, data for the top 20 virtual disks selected by the highest Average Latency is displayed. Virtual disks are color-coded according to the key at the top of the chart.
By default, the virtual disks are sorted by the selected statistic, but the set of virtual disks can be sorted by any one of the statistics provided in the tool. The bars representing the virtual disks in each chart will be ordered according to the selected sorting statistic from left to right. The virtual disks in all three charts are always in alignment based on the selected sort order.
A tool tip defines the statistic when the mouse is hovered over a bar in the chart. The tool tip identifies the virtual disk and the specific value for the statistic. If the virtual disk name appears in italics, data is incomplete for the selected time period.
The table view provides more detailed information and presents performance statistics for all served virtual disks in the server group in a table format. Under each virtual disk are individual values per host. The value displayed per host in the table view shows the statistics for paths between the host and the virtual disk. In addition, DataCore Servers owning the storage sources in each virtual disk are listed in the table for convenience.
Columns in the table are sortable. By default, column data is sorted by Average Latency from highest to lowest values, but data can be sorted by any column so that administrators can quickly determine virtual disks with the highest values in order to troubleshoot performance issues. Data is sorted by one column at a time.
Column cells are color-coded, to match the bar chart, with highest values having the darkest color to quickly determine the highest values in the list.
A tool tip defines the statistic when the mouse is hovered over the heading in the table.
To sort data in columns:
o The arrow indicates the column by which the table has been sorted. Click the column header to change the column by which the table is sorted,
o A down-pointing arrow means that values are sorted from high to low. An up-pointing arrow means that values are sorted from low to high. Change the arrow direction by clicking again in the column header.
To filter data in columns:
o The funnel symbol in
the column header removes all data that does not meet the selected value.
Multiple columns can be simultaneously filtered to exclude data that does
not meet all selected values. To select a value, focus the mouse in a
column and click the funnel symbol, then select the value from the drop-down
menu. Active filters are displayed at the bottom of the table and can
be disabled by clearing the check box or removed by clicking X.
All tool selections appear on the left side of the tab.
o The top 20 virtual disks to display and the sort order can be selected in the Chart view by clicking a radio button to the left of the charts.
o The time period to display can be selected in either view by clicking a radio button on the left side of the tab.
o The view can be toggled between Chart and Table view by clicking the link Table View or Chart View in the upper-left corner of the tab.
The data currently displayed in the list view of the tool can be exported. This includes all virtual disk data and data displayed per host. Virtual Disk IDs and Host IDs are also included.
To export data to a CSV file:
1 In the Performance Spotlight, select the time period for the data desired in the CSV file.
2 In the System Health Actions tab in the Ribbon, click Export to CSV.
3 In the Save As dialog box, browse to the folder where the file should be saved. Keep the default name provided in the dialog box or enter a different file name.
4 Click Save to save the data to a file.
