Thursday Sep 26, 2013

Bounce rate is not everyone's enemy

I am cross-posting here an interesting piece of info & advice that my colleague Abhijith Ramalingaiah shared on an Oracle alias then on his slideshare page. It has been a while since I have posted something about web analytics, a topic you quickly get interested in after starting your own blog, thanks to Abhijith for reviving it :-)

Abhijith is taking a closer look at Bounce Rate (BR). BR represents the percentage of visitors who enter the site and leave ("bounce") rather than continue viewing other pages on the same site. High BR is thus typically viewed as a sign of poor performance, visitors are leaving. Abhijith is telling us however that it is not necessarily such a negative sign, and that the content on the page has to be considered as well as visitor experience.

Abhijith then goes thru the following scenarios where BR is not a good measurement for the page.

[Read More]

Thursday Sep 05, 2013

Misys Kondor+ runs best on SPARC T5

Misys is a leading financial software vendor, providing the broadest portfolio of banking, treasury, trading and risk solutions available on the market. At ISV Engineering, we have a long-standing collaboration with the Kondor+ product line, that came from the Turaz acquisition (formerly part of Thomson Reuters) and now part of the Misys Treasury & Capital Markets (TCM) portfolio. Jean-Marc Jacquot, Senior Technical Consultant at Misys TCM, was recently interviewed by ITplace.TV (in French language) about a recent IT redesign of the Kondor+ installed base at a major French financial institution.

The customer was running various releases of Kondor+ over a wide range of SPARC machines, from Sun V240 to M4000. The IT project aimed at rationalizing these deployments for better ROI, SLA and raw performance —to meet the new system requirements of the latest Kondor+ release. In the short list, SPARC & Solaris beat x86 & Linux on …

[Read More]

Friday Aug 23, 2013

OpenWorld and JavaOne 2013 Sessions from ISV Engineering

Come meet members of Oracle's ISV Engineering during OpenWorld and JavaOne this year. Here is a list of sessions where we will be speaking at. See you there!

JavaOne BOF4681 Quick Ad Hoc Instrumentation of DTrace Probes by a VisualVM Plug-in Amit Hurvitz
OpenWorld CON5841 Oracle Exa-stack: Scalable, Secure, Performant Core Banking with Temenos T24 Matt Hosanee
OpenWorld CON8635 Accelerate SAP with Oracle SuperCluster Hans-Juergen Denecke
OpenWorld HOL10181 Oracle Solaris Integrated Load Balancer in 60 Minutes Amir Javanshir
OpenWorld HOL10182 How to Set Up a Hadoop Cluster with Oracle Solaris Orgad Kimchi
OpenWorld HOL10194 Using DTrace to Troubleshoot Java Native Memory Problems on Oracle Solaris 11 Gang Wang
Wen-Sheng Liu
Yu Wang

Tuesday Jun 18, 2013

Moving Oracle Solaris 11 Zones between physical servers

As part of my job in the ISV Engineering team, I am often asked by partners the following question : is it possible to easily move a Solaris 11 Zone from a physical server to another?

The short answer is : YES ! The longer one comes with the following restrictions :

  • Both physical servers should be of the same architecture, x64 or SPARC (T-series and M-series systems are compatible).

  • Both physical servers should run Oracle Solaris 11.

  • The destination server should run at least the same or higher release of Solaris 11. This includes the SRU (Support Repository Update) level.

Given a physical server called “Source” hosting a Solaris 11 Zone called “myZone” on a ZFS filesystem, here are the steps to move the zone on another physical server called “Target” :
  1. Export the Zones configuration
    The zone needs to be configured on the destination server before it can be installed. The first step is to export the configuration of the Zone to a file:

    [Source]# zonecfg -z myZone export -f myZone.cfg

  2. Archive the Zone
    My favourite solution is to use the ZFS “send” functionality to archive the ZFS file system hosting the Zone in a single movable file, although this can also be achieved in other ways (cpio, pax)

    • Halt the Zone
      [Source]# zoneadm -z myZone halt

    • Take a recursive ZFS Snapshot of the rpool of the zone
      [Source]# zfs snapshot -r rpool/zones/myZone@archive

    • Archive the Zone using ZFS send (ZFS and cpio archives can be zipped using gzip or bzip2)
      [Source]# zfs send -rc rpool/zones/myZone@archive | bzip2 > /var/tmp/myZone.zfs.bz2
  1. Move the configuration and archive files to the destination server
    FTP, scp, NFS, removable hard drive, …

  2. Configure the zone on the destination server

    Depending on the configuration of the Targer server, you might need to tweak the zone configuration file before using it.

    [Target]# zonecfg -z myZone -f myZone.cfg

  3. Install the Zone

    If the zone is being installed in the same network, the zone configuration (IP address, DNS server, Gateway, etc) can be preserved using the "-p" option:

    [Target]# zoneadm -z myZone install -a myZone.zfs.bz2 -p

    If the Zone is being installed in a new network environment, using the "-u" option instead of "-p" will unconfigure the system. The Zone would need to be manually configured on the first boot. The configuration can be automatized during installation if a system configuration profile XML file is provided:

    # zoneadm -z myZone install -a myZone.zfs -u -c sc_profile.xml

    Quick Tip
    : To create a system configuration file, you can use the sysconfig program with option "create-profile":

    # sysconfig create-profile -o sc_profile.xml

    The configuration text wizard will walk you through the system configuration steps (same process as the first boot configuration wizard) but will not re-configure your system. It will simply create an output XML file with the defined system configuration. This files can then be manually tweaked if needed and act as a template for future use.

  4. Boot the Zone
    # zoneadm -z myZone boot

  5. You should now be able to log in the Zone which is the exact copy of the original Zone on the source server.

Obviously there are many more options and possibilities that go beyond the scope of this post. My intend was just to give a glimpse of what can be done, so don't hesitate to consult the documentation for more options.

Also, these simple steps cans be scripted to be made even more flexible and usable. Below are two scripts I have written for my own needs. There are only provided as an example and must not be considered as production ready scripts.





# This script creates a movable archive of an Solaris 11 Zone

# It take a single input parameter: The Zone name


BASE_DIR="$(pwd -P)"





SNAPSHOT=${ZONES_ROOT}/${ZONE_NAME}@`date '+%d_%m_%Y-%H:%M:%S'`

if [ if [ ! -d ${ARCHIVE_FOLDER} ] ; then
   mkdir -p ${ARCHIVE_FOLDER}

zoneadm -z ${ZONE_NAME} halt

zonecfg -z ${ZONE_NAME} export -f ${ARCHIVE_FOLDER}/${CONFIG_FILE}

# Take a ZFS Snapshot of the rpool of the zone
zfs snapshot -r ${SNAPSHOT}

# Archive the Zone using ZFS send
zfs send -rc ${SNAPSHOT} | bzip2 > ${ARCHIVE_FOLDER}/${ARCHIVE_FILE}

# Delete the snapshot used to create the archive
zfs destroy -r ${SNAPSHOT}





# This script deploys an archived Solaris 11 Zone

# It take a single input parameter: The Zone name



BASE_DIR="$(pwd -P)"




# Configure the Zone

# Install the Zone and configure the system
zoneadm -z ${ZONE_NAME} install -a ${ARCHIVE_FOLDER}/${ARCHIVE_FILE} -u

# Boot the Zone
zoneadm -z ${ZONE_NAME} boot

Friday Jun 14, 2013

Presentations from IGTCloud Meetup, June 6th, Herzliya

Thanks for joining us for the IGTCloud Meetup last week! I am posting here presentations from the event. I am also pointing you to the Oracle Cloud Solutions website where you will find current and exhaustive product information. Please continue to send feedback to help us improve these Tech Days.


How open innovation and technology adoption translates to business value, with stories from our developer support work at Oracle's ISV Engineering.



« April 2014