Highlighted
Absent Member.
Absent Member.
85 views

How to measure round trip between ucmdb and dbms server

Hi,
we are in uCMDB 902 + Content Pack 8 under Windows 2008.
Could you confirm me that access to database used to store data is based on jdbc driver ?
In order to explain difference of load time for 7000 ci between two environments, we want to measure round trip between ucmdb server and dbms server (MS SQL SERVER 2008).
In uCMDB directory or with third party tool, how to measure round-trip time ?
Thanks.

Jean-Marc Friaud
0 Likes
4 Replies
Highlighted
Micro Focus Contributor
Micro Focus Contributor

Re: How to measure round trip between ucmdb and dbms server

Hi

Not sure why you want to calculate RTT as the application does not send same data back and forth. I am assuming you want to know how to check why inserting same 7000 CIs are taking longer in one env as compare to another similar UCMDB env ( Same Application server hardware and configuration, same network latency and same DB configuration).

If so, you can check cmdb.reconciliation.log:

 

It will explain how much time it took to insert data: 

 

2013-01-11 08:30:59,431 INFO  -  [ID= 654844694] [total time= 3.851 model= 0.136] addOrUpdate    :  [durations: identify= 0.212 dataIn= 0.059] Ignore From Cmdb-  0 Merge Operations-  0 Merged CIs-    0 Type Changes-    0 Max Topology Level- 1]

 

dataIn is the DB time. The rest is the time taken by CMDB.

 

Also HP has tool called "load runner" that in general is used for load testing.

 

0 Likes
Highlighted
Absent Member.
Absent Member.

Re: How to measure round trip between ucmdb and dbms server

Hi,

please find attached our cmdb.reconciliation.log. We don't have entries durations in this file.

To explain differences of insert time between two environment, i only need a way or a tool to trace a SQL round-trip message, first byte from ucmdb server to database server, and last byte of result.

Could you help me ?

 

Jean-Marc Friaud
0 Likes
Highlighted
Super Contributor.
Super Contributor.

Re: How to measure round trip between ucmdb and dbms server

Hi,

As you mentioned, data access is being done via jdbc drivers, which are Mercury(HP)-branded DataDirect drivers. They offer quite good performance if your db environment has sufficient capacity, as described in Database Guide.

Tracing SQL commands would be tedious and wouldn't give usefull information. uCMDB application stores part of the data model in-memory, so db operations are not necessairly issued right after data in. Besides, uCMDB app uses ORM mapping (hibernate), which generates sql queries, so they're hard to read. If you're really interested in SQL statistics, than increase logging levels for cmdb.dal.* logs and observe them.

 

Overall performance of data-in operations are dependent on the reconciliation mechanism. If your input bulk does not include sufficient identification data and in uCMDB there are lots of generic CIs (like host or host_node), than identification maps can grow exponentialy. This could even lead to total freeze of uCMDB.

 

First place to look is cmdb.reconciliation.audit.log. It gives good overview of performance of each data input operation. If you're curious about performance of individual job, than you should check other cmdb.reconciliation logs increasing logging level if necessary. You could also use adapters' "Include results in communication log", to observe what data is being sent to server.

 

kind regards

Grzegorz Kluczek/

0 Likes
Highlighted
Absent Member.
Absent Member.

Re: How to measure round trip between ucmdb and dbms server

Hi Grzegorz,

I need to undestand difference of performance between two environnements for inserting 7000 ci with Connect-IT scenario.

An hypothesis is round trip differences between two database servers.

You indicate HP uCMDB is using Mercury(HP)-branded DataDirect. On Datadirect web site, there is a troubleshooting tool called snoop

http://www.datadirect.com/support/downloads/tools.html

 

Could you simply help me to use snoop in a HP uCMDB contexte ?

We are using MS SQL 2008 for database access.

 

Regards.

 

Jean-Marc Friaud
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.