Re:Repository Adapter

[vkaranam]

Hello Guys,

We are on vcops 5.7 and i have configured the repository adapter for the 5.7. It is up and running and collecting the data from the vcops. The problem is it is collecting the repeated data for all the machines present. The CSV file is of 2GB size just for 150 VM's. The following are the metrics that are defined in conf.properties file.

insertCommand = INSERT INTO TestTable (RID1; RNAME1; MID1; MNAME1; RKNAME1; MKNAME1; AKNAME1; timestamp1; min_threshold1; value1; max_threshold1) values(alive.RID; alive.RNAME; alive.MID; alive.MNAME; alive.RKNAME; alive.MKNAME; alive.AKNAME; fsdb.timestamp; fsdb.min_threshold; fsdb.value; fsdb.max_threshold)

In the conditions.properties i have defined these metrics.

RKNAME1='HostSystem'

MNAME1=<>'cpu*';'datastore*';Disk*';'mem*'

RID1

RNAME1

MID1

MNAME1

RKNAME1

MKNAME1

timestamp1

min_threshold1

value1

max_threshold1

Host 1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	0	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	0	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	0	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	4.666666667	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	0	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	4.666666667	14.0020256
Host1		1252	storageAdapter:vmhba0|read_average	HostSystem	read_average	1.36656E+12	0	0	14.0020256

The above pasted is a sample excel file with few columns. For the host 1 itself i got 104746 colums of data with related to cpu, datastore disk and memory. In that all the data is mostly repeated. i am unable to figure out how to filter the data and get the sharp values for all the hosts. say for example i have 10 hosts so 10*4(cpu,datastore,disk,mem) = 40 columns. how can i reduce that data??

Also any ideas how your using it in ur environment.

Any help is appreciated.

Thanks

VK

[TammieD]

I haven't done this, but was looking into it because I was looking to do something similar.

Based on what I've read, if your metric is collecting every 5 minutes that is 288 data points for one metric for one resource per day.  You can collect just 1 day of data using the maxDays in the conf.properties file.  I don't see anything that would select just the latest value for a metric for a resource.

However, you also have a wild card in the metric names, which is going to give you more than just 4.  You would need to get the exact metric name or metric ID for the ones you want.

[VirtualBhas]

Rest assured the data are not repeats tough they look like it. The timestamp column is in the Epoch / Unix time format. You can use this formula in Excel to convert it to human readable format. vCOPs time has ms as well, so remember to trim the last 3 digits off if you are using the below formula in excel.

=TEXT(A1/(60*60*24)+"1/1/1970","yyyy/mm/dd hh:mm:ss")

This will the Eastern Time

Keep in mind, vCOPs collects data every 5 minutes. So if you are generating the CSV for a day, you will see 288 (12*24=288) rows for each metric for each host. You have 10 Hosts that is 288*10 = 2880 rows per metric. Since you are saying MNAME1=<>'cpu*';'datastore*';Disk*';'mem*', it will generate for all the metrics under CPU, Datastore, Disk and mem. That's the reason your CSV is so big.

Try to narrow down the metrics you want in the csv in the conditions.properties file

Eg:

MNAME1='cpu|usage_average';'mem|host_usagePct'

This will generate the CSV only for 2 metrics Average CPU Usage and Host Memory Usage (%)

You can get the list of all metrics from the vCenter Adapter Install and Config Guide

ftp://ftp.integrien.com/VCOPS-VMware_Product_Adapters/VMware-vCenter/vCenter-Adapter-Install-and-Con...

All the Best.