Upgrading from SM9.21 to 9.4x - time estimate

I do realize this is a question that cannot be answered with certanity....but surely there are some customers who have upgraded from pre 9.3 to 9.4x.  If I could just get a feel for their experience and/or talk with them, that would be great.

I need a gernade range estimate of man hours and calendar time required, based on past experience, to upgrade from 9.21 (or earlier) to 9.41 codeless, given that we have medium amount of customizations that we will need to convert and or use in a hybrid environment

Assume that man hours are coding and configuration "seat time", performed by SM "experts".

We have integrations to/from the following:

    • IIQ?
    • BSM / OMI
    • Bomgar
    • Maximo
    • HP Asset Manager
    • HP uCMDB
    • Microsoft SCCM
  • Additionally we have multiple small integrations using SOAP API and/or WebServices

Respectfully

Stacy Whetzell

  • Greetings.

    I don't think that even an estimate time can be given. There are a lot of things to review like data, customization fields and of course the integrations. My suggestion is to open a case with the support team and that way it can be contacted to the upgrade factory so that way they can check all that is required for the upgrade and the estimated time for the tasks to be performed.

  • The upgrade time is directly proportional to the customization done. The time to upgrade increases if you are using a highly customized environment and trying to use new functionality and replace old tailoring. The upgrade time does not impact your production as the upgrade is done in a test environment and the package "custom upgrade" is then applied to your production.

    you will need two machine

    Machine A

    1) restore production database

    2)download upgrade utility

    3) resolve conflict resolution

    4)take a backup point A

    5) create custom upgrade

    Machine B

    1) restore production database

    2) Apply custom upgrade

    this will be the downtime in production ( can be somewhere between 4-6 hrs)

     

  • I'm starting to form what is likely a premature opinon based on responses to my many posts today.

    Given:

    - We have a heavily customized 9.21 environment.

    - A minimum of 2 upgrades will be required for us to acutally 'upgrade' from 9.21 to 9.41 codeless or hybrid

    - Some, or perhaps even 'many', of our customizations may be OOTB functionality in 9.41

    - Upgrading requires extensive testing to ensure "all" code relate to a particular customization have been moved up and recoded (if necessary) to make compliant with 9.41.

    - Probability that "happy path" acceptance testing will, in spite of best intentions, result in some production breakages of processes in days and weeks following "go-live" of an upgraded system

    - Our current customizations include some 'garbage' that we likely wish to remove

    - Our acutal processes, though documented poorly, are pretty widely known and accepted within the development team

    My premature inclination is that if Management is willing to accept a swivel seat, two enviornment solution for a few months (to complete/close tickets in old system), we may be better off in standing up an OOTB installation of 9.4x codeless and rebuilding all our processes from the beginning.

    I assume this approach would also mean that we have to:

    -  leave legacy 'ticket' information behind in the old system (to be archieved at some point in time)

    What else might we be loosing with the OOTB 9.41 codeless, rebuild all from scratch solution?

    Also, does my premature inclination seem to be bad assumptions? or are there others who agree that starting with a 'clean slate' might be best?

    thank you so much

    Stacy Whetzell

     

     

     

  • Hi Stacy,

    Something about estimated conflict resolution effort you can refer to you other post - http://community.hpe.com/t5/Service-Manager-Support-Customer/SM9-4x-migration-upgrade-tools/m-p/6829650

    For your legacy tickets, if you can, better to keep them running in old version system, and create new ticket in new version system.

    If you have clean process and tailoring code, even they are heavily customized, you are recommended to use strategy of reimplementation on new installed 9.41. You can oboselete some tailoring codes which have been implemented in out of box. You can migrate your master data, like operator/contacts/group/device/assignment/etc via database tools/excel/HPE Connect-It/SM9.41 Delta Migration tool, even your developed scripts if you are familar the table/fields.

    Note: You can NOT enable Service Desk & Incident module by module in PD mode in this way, but you will be in forcely full Codeless mode.

    Pitt,

    SM RnD Upgrade Factory

  • Hi Stacy,

    As a reference, another customer who have also more than 5 integrations with SM, and medium level customization, planned to upgrade to 9.41 Hybrid, and estimated their total upgrade project will take about 120 man-days. Actually most of time (~50%) are on testing (several cycles of function testing, system integration testing, user acceptable testing, load testing), ~25% on development (running upgrade/conflicts resolution/bug fixing), ~25% on preparation/plan/meeting/training/misc...

    Pitt,

    SM RnD Upgrade Factory

  • Hi Stacy,

    As a reference, another customer who have also more than 5 integrations with SM, and medium level customization, planned to upgrade to 9.41 Hybrid, and estimated their total upgrade project will take about 120 man-days. Actually most of time (~50%) are on testing (several cycles of function testing, system integration testing, user acceptable testing, load testing), ~25% on development (running upgrade/conflicts resolution/bug fixing), ~25% on preparation/plan/meeting/training/misc...

    Pitt,

    SM RnD Upgrade Factory

  • hi Stacy,

    Please read the reply from Pitt, he is part of the upgrade factory. For the question "Also, does my premature inclination seem to be bad assumptions? or are there others who agree that starting with a 'clean slate' might be best?"

    I would think the best approach will be for you to install SM9.41 oob and compare all of the feature with your existing SM9.21 code and  document all of the tailoring in SM9.21 that you can do away in SM9.41.

    Best Scenario: you can do a fresh implementation of SM9.41 in production and have SM9.21 along side till your business can decommision SM9.21 and use SM9.41 as production db.

    another Scenario: You check all of the old tailor code that the SM9.41 feature replaces and proceed with an upgrade to SM9.41 deciding of you want to install PDCP before or after i.e decide SM9.41 codeless or hybrid

    Let us know how can we be of assistance to you.