We hope you'll join us for our 4/23 webinar on using data tables to apply reference ranges and AE codes in OC4. For more information and to register, visit https://register.gotowebinar.com/register/2882170018956684555

OpenClinica Slowness

2»

Comments

  • jamuna269jamuna269 Posts: 109
    Mihai,
    I see time stamp along with date_created field and it looks like it got upgraded to include timestamp in 3.1+ versions.
    Thanks
    Jamuna
    On Fri, Feb 22, 2013 at 2:16 PM, Mihai Virtosu wrote:
    Thanks Jamuna!

    May I suggest adding a time component to “date_created”?

    Mihai
    Sent: Friday, February 22, 2013 12:05 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness

    Hi Mihai,
    There is a table called archived_file_dataset and it should have the information about which file was extracted when.
    Thanks
    Jamuna

    On Fri, Feb 22, 2013 at 11:56 AM, Mihai Virtosu wrote:
    Richard,
    We are still using 3.0.4.2 and hence do not have a choice of log level.
    Thanks,
    Mihai
    -----Original Message-----
    Sent: Thursday, February 21, 2013 2:24 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness
    Hi,
    Could it be caused by the logging level chosen?
    Richard
    On 22/02/13 6:56 AM, "Gerben Rienk" wrote:
    >Hello Mihai,
    >
    >That's quite some hardware ;-)
    >You did not mention the settings for tomcat and for postgres.
    >If you did not tweak postgres, try
    >effective_cahce_size=2GB
    >Kind regards,
    >
    >Gerben Rienk
    >
    >> Hello OC users,
    >>
    >> Has anyone run in the problem of OpenClinica becoming very slow when
    >>many users are using the system, and if yes, what was your solution,
    >>from a technical perspective, other than reducing CRF size, limiting
    >>number of rules and so on?
    >>
    >> Our OC specs are below:
    >>
    >> VMware ESX 4.1 machine
    >> OS: Windows 2003 SP2 32 bit
    >> Processor: Intel Xeon x5680 3.33 GHz running 4 cores
    >> Ram: 4 GB DDR3
    >> HDD: Equalogic SAN with SAS Harddrives on RAID 10
    >> Network: 10Gb to SAN 1 Gb to LAN
    >>
    >> Thank you,
    >>
    >> Mihai Virtosu
    >> University of Utah
    >>
    >>
  • zwets-kcrizwets-kcri Posts: 50
    May,
    > > Where should I add this line for postgres configuration?
    > >
    >> >> If you did not tweak postgres, try
    >> >> effective_cahce_size=2GB
    Apart from this setting, there is a lot more information on tuning PostgreSQL at: http://wiki.postgresql.org/wiki/Tuning_Your_PostgreSQL_Server
    Since PostgreSQL is the bottleneck of the application, and as Gerben mentioned, is configured for tiny machines by default, it is a must to go over these.
    As for Tomcat, the only relevant parameter to tune is the amount of JVM heap allocated. You do this with the -Xmx parameter which you set in the CATALINA_OPTS variable in setenv.sh (Linux, MacOS) or setenv.bat (Windows). Start out with e.g. -Xmx=800m and monitor the actual heap usage with jconsole or using the online probe at www.java-monitor.com. Set it so that usage doesn't regularly exceed the 70-80% mark. Don't set -Xmx excessively high either (thinking "I've got plenty of memory anyway") because that will hurt performance too.
    If you do frequent reloads/restarts of web applications without restarting the Tomcat instance, then you will eventually run into "Out of PermGen memory" errors, and you will need to set the -XX:MaxPermSize parameter. A suggested value is -XX:MaxPermSize=120m, but jconsole a/o java-monitor will give you an empirical evidence.
    Cheers
    Marco
    --
    Marco van Zwetselaar
    KCRI - Kilimanjaro Clinical Research Institute
    Moshi
    Tanzania
    e [email protected]
    m +255 782 334124
  • mvirtosumvirtosu Posts: 276 ✭✭
    Jamuna,

    What is the measurement unit for “run_time” in this table? Has it changed since version 3.0.4.2?

    Thanks,

    Mihai
    Sent: Friday, February 22, 2013 2:26 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness

    Mihai,
    I see time stamp along with date_created field and it looks like it got upgraded to include timestamp in 3.1+ versions.
    Thanks
    Jamuna

    On Fri, Feb 22, 2013 at 2:16 PM, Mihai Virtosu wrote:
    Thanks Jamuna!

    May I suggest adding a time component to “date_created”?

    Mihai
    Sent: Friday, February 22, 2013 12:05 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness

    Hi Mihai,
    There is a table called archived_file_dataset and it should have the information about which file was extracted when.
    Thanks
    Jamuna

    On Fri, Feb 22, 2013 at 11:56 AM, Mihai Virtosu wrote:
    Richard,
    We are still using 3.0.4.2 and hence do not have a choice of log level.
    Thanks,
    Mihai
    -----Original Message-----
    Sent: Thursday, February 21, 2013 2:24 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness
    Hi,
    Could it be caused by the logging level chosen?
    Richard
    On 22/02/13 6:56 AM, "Gerben Rienk" wrote:
    >Hello Mihai,
    >
    >That's quite some hardware ;-)
    >You did not mention the settings for tomcat and for postgres.
    >If you did not tweak postgres, try
    >effective_cahce_size=2GB
    >Kind regards,
    >
    >Gerben Rienk
    >
    >> Hello OC users,
    >>
    >> Has anyone run in the problem of OpenClinica becoming very slow when
    >>many users are using the system, and if yes, what was your solution,
    >>from a technical perspective, other than reducing CRF size, limiting
    >>number of rules and so on?
    >>
    >> Our OC specs are below:
    >>
    >> VMware ESX 4.1 machine
    >> OS: Windows 2003 SP2 32 bit
    >> Processor: Intel Xeon x5680 3.33 GHz running 4 cores
    >> Ram: 4 GB DDR3
    >> HDD: Equalogic SAN with SAS Harddrives on RAID 10
    >> Network: 10Gb to SAN 1 Gb to LAN
    >>
    >> Thank you,
    >>
    >> Mihai Virtosu
    >> University of Utah
    >>
    >>
  • jamuna269jamuna269 Posts: 109
    Mihai,
    I believe the way we calculate the field has been changed to reflect the accurate value.
    Jamuna
    On Fri, Mar 8, 2013 at 2:03 PM, Mihai Virtosu wrote:
    Jamuna,

    What is the measurement unit for “run_time” in this table? Has it changed since version 3.0.4.2?

    Thanks,

    Mihai
    Sent: Friday, February 22, 2013 2:26 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness

    Mihai,
    I see time stamp along with date_created field and it looks like it got upgraded to include timestamp in 3.1+ versions.
    Thanks
    Jamuna

    On Fri, Feb 22, 2013 at 2:16 PM, Mihai Virtosu wrote:
    Thanks Jamuna!

    May I suggest adding a time component to “date_created”?

    Mihai
    Sent: Friday, February 22, 2013 12:05 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness

    Hi Mihai,
    There is a table called archived_file_dataset and it should have the information about which file was extracted when.
    Thanks
    Jamuna

    On Fri, Feb 22, 2013 at 11:56 AM, Mihai Virtosu wrote:
    Richard,
    We are still using 3.0.4.2 and hence do not have a choice of log level.
    Thanks,
    Mihai
    -----Original Message-----
    Sent: Thursday, February 21, 2013 2:24 PM
    To: [email protected]
    Subject: Re: [Users] OpenClinica Slowness
    Hi,
    Could it be caused by the logging level chosen?
    Richard
    On 22/02/13 6:56 AM, "Gerben Rienk" wrote:
    >Hello Mihai,
    >
    >That's quite some hardware ;-)
    >You did not mention the settings for tomcat and for postgres.
    >If you did not tweak postgres, try
    >effective_cahce_size=2GB
    >Kind regards,
    >
    >Gerben Rienk
    >
    >> Hello OC users,
    >>
    >> Has anyone run in the problem of OpenClinica becoming very slow when
    >>many users are using the system, and if yes, what was your solution,
    >>from a technical perspective, other than reducing CRF size, limiting
    >>number of rules and so on?
    >>
    >> Our OC specs are below:
    >>
    >> VMware ESX 4.1 machine
    >> OS: Windows 2003 SP2 32 bit
    >> Processor: Intel Xeon x5680 3.33 GHz running 4 cores
    >> Ram: 4 GB DDR3
    >> HDD: Equalogic SAN with SAS Harddrives on RAID 10
    >> Network: 10Gb to SAN 1 Gb to LAN
    >>
    >> Thank you,
    >>
    >> Mihai Virtosu
    >> University of Utah
    >>
    >>
This discussion has been closed.