I have a number of customers who want to run staggered deduplication scedules through the night, or spread over the week, or both. To address this requirement in a workflow, I wrote this simple function which pseudorandomly selects a day@hour string for deduplication schedules. As is, it selects any day of the week and a time between 22:00 - 5:00, but this can be easily customized to different customer requirements.
The idea can of course be expanded to other "random" scheduling tasks (reallocate...).
Message was edited by: Jonathan Tourtellot - replaced DAR with cleaned-up version
Could someone just post the ASCII text of this function if they have it? This DAR won't import into WFA-2.2. I can peek inside DAR (.zip) and see the code but it's distorted by being in XML form that way. It needs some manual editing.
Re: A function to generate a deduplication schedule
I just came across this post, and figured I will throw my version into the mix. My version is actually a command and not a function. It's based on Cron jobs already being setup with a naming convention of "maint_##_r" for replication, and "maint_##_d" for deduplication. The "_d" will run an hour before the "_r"(see below). It will choose the appropriate "_r" job based on the least number of snapmirror relationships using those cron schedules. You can easily take the script and turn it into one that looks at the capacity of what is replicated as well, but I chose not to look at that as it isn't a true measurement of the change rate.
Once it chooses the replication cron job, it will also choose the associated deduplication job, and add those variables into the workflow parameters for later use down the line of the workflow.
It also checks to see if the efficiency schedule is already setup on the source side, and creates the effeciency policy if it doesn't exist.