We are going to be doing some testing of time sensitive data using RHQ. We want to run successive tests with a internal NTP server making the servers think they are always in the same time period. That is, we will reset the NTP clock back before each test. This is because the data being tested has business logic that relates to the current time.
We want to purge the RHQ server of data from each previous test so the identically-timestamped data don't collide. Is this viable? Would it be a matter of truncating certain tables? Or would it be better to snapshot the whole machine?
Thanks, John
If you have the source code, you can run dbsetup to reset your database. git checkout the source, make sure your maven settings.xml is setup properly (see etc/m2/settinsg.xml in the source for an example) then:
cd modules/core/dbutils mvn -Ddbsetup -Dmaven.test.skip install
This will drop all tables, then recreate them from scratch - you'll lose all inventory and data when you do this - you start as if you just installed the RHQ Server anew. Make sure you stop the RHQ Server before doing this, obviously. After dbsetup finishes (takes a few seconds), restart the RHQ Server.
If you don't want to bother going thru this setup of the build from source, just do an initial install of your RHQ Server (i.e. unzip the distro, run it, go through the UI installer to install the server), then backup the DB. When you want to restart, just stop the RHQ Server, revert your DB backup, then restart the RHQ Server. I do this alot on Postgres - its really easy to backup and restore on Postgres using their admin tool.
On 05/05/2011 11:59 AM, John Hollland wrote:
We are going to be doing some testing of time sensitive data using RHQ. We want to run successive tests with a internal NTP server making the servers think they are always in the same time period. That is, we will reset the NTP clock back before each test. This is because the data being tested has business logic that relates to the current time.
We want to purge the RHQ server of data from each previous test so the identically-timestamped data don't collide. Is this viable? Would it be a matter of truncating certain tables? Or would it be better to snapshot the whole machine?
John, Thanks for all the information!
John
On 5/5/2011 1:02 PM, John Mazzitelli wrote:
If you have the source code, you can run dbsetup to reset your database. git checkout the source, make sure your maven settings.xml is setup properly (see etc/m2/settinsg.xml in the source for an example) then:
cd modules/core/dbutils mvn -Ddbsetup -Dmaven.test.skip install
This will drop all tables, then recreate them from scratch - you'll lose all inventory and data when you do this - you start as if you just installed the RHQ Server anew. Make sure you stop the RHQ Server before doing this, obviously. After dbsetup finishes (takes a few seconds), restart the RHQ Server.
If you don't want to bother going thru this setup of the build from source, just do an initial install of your RHQ Server (i.e. unzip the distro, run it, go through the UI installer to install the server), then backup the DB. When you want to restart, just stop the RHQ Server, revert your DB backup, then restart the RHQ Server. I do this alot on Postgres - its really easy to backup and restore on Postgres using their admin tool.
On 05/05/2011 11:59 AM, John Hollland wrote:
We are going to be doing some testing of time sensitive data using RHQ. We want to run successive tests with a internal NTP server making the servers think they are always in the same time period. That is, we will reset the NTP clock back before each test. This is because the data being tested has business logic that relates to the current time.
We want to purge the RHQ server of data from each previous test so the identically-timestamped data don't collide. Is this viable? Would it be a matter of truncating certain tables? Or would it be better to snapshot the whole machine?
rhq-users mailing list rhq-users@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/rhq-users
For the record, there is a wiki page that contains info on the dbsetup stuff... see:
http://rhq-project.org/display/RHQ/Advanced+Build+Notes#AdvancedBuildNotes-P...
On 05/05/2011 01:02 PM, John Mazzitelli wrote:
If you have the source code, you can run dbsetup to reset your database. git checkout the source, make sure your maven settings.xml is setup properly (see etc/m2/settinsg.xml in the source for an example) then:
cd modules/core/dbutils mvn -Ddbsetup -Dmaven.test.skip install
This will drop all tables, then recreate them from scratch - you'll lose all inventory and data when you do this - you start as if you just installed the RHQ Server anew. Make sure you stop the RHQ Server before doing this, obviously. After dbsetup finishes (takes a few seconds), restart the RHQ Server.
If you don't want to bother going thru this setup of the build from source, just do an initial install of your RHQ Server (i.e. unzip the distro, run it, go through the UI installer to install the server), then backup the DB. When you want to restart, just stop the RHQ Server, revert your DB backup, then restart the RHQ Server. I do this alot on Postgres - its really easy to backup and restore on Postgres using their admin tool.
On 05/05/2011 11:59 AM, John Hollland wrote:
We are going to be doing some testing of time sensitive data using RHQ. We want to run successive tests with a internal NTP server making the servers think they are always in the same time period. That is, we will reset the NTP clock back before each test. This is because the data being tested has business logic that relates to the current time.
We want to purge the RHQ server of data from each previous test so the identically-timestamped data don't collide. Is this viable? Would it be a matter of truncating certain tables? Or would it be better to snapshot the whole machine?
rhq-users mailing list rhq-users@lists.fedorahosted.org https://fedorahosted.org/mailman/listinfo/rhq-users
On 05/05/2011 01:02 PM, John Mazzitelli wrote:
If you have the source code, you can run dbsetup to reset your database. git checkout the source, make sure your maven settings.xml is setup properly (see etc/m2/settinsg.xml in the source for an example) then:
cd modules/core/dbutils mvn -Ddbsetup -Dmaven.test.skip install
This will drop all tables, then recreate them from scratch - you'll lose all inventory and data when you do this - you start as if you just installed the RHQ Server anew. Make sure you stop the RHQ Server before doing this, obviously. After dbsetup finishes (takes a few seconds), restart the RHQ Server.
If you don't want to bother going thru this setup of the build from source, just do an initial install of your RHQ Server (i.e. unzip the distro, run it, go through the UI installer to install the server),
Note, assuming this is a reinstall, make sure to choose "Overwrite existing data" in the installer, rather than "Upgrade existing data". The overwrite option is the equivalent of running the dbsetup utility.
then backup the DB. When you want to restart, just stop the RHQ Server, revert your DB backup, then restart the RHQ Server. I do this alot on Postgres - its really easy to backup and restore on Postgres using their admin tool.
rhq-users@lists.fedorahosted.org