This is how it works for me:
- IT states they're going to do updating on a certain date.
- I send out an email to my licensed Tableau people letting them know when the operation will be happening so they won't be using the server during that time.
- IT does their work and reboots the server.
- That morning I check that my Tableau Server is running.
As far as I know, there is no need for you to stop/start Tableau services nor stop/start the Tableau Server. Tableau Server is easy!
Even though it's possible to have Tableau Services come back up with the Windows Box after the reboots complete, wouldn't it be best practices to perform a tabadmin stop so that things like the Data Engine aren't just "cut off" and they are properly brought down from the command?
Hi Toby ,
any thoughts over Brian's reply to this thread ?
"wouldn't be best practices to perform a tabadmin stop so that things like the data engine aren't just "cut off" and they are properly brought down from the commad "
First, my pet-peeve and this isn't directed directly at anyone, I don't like the term "Best Practices". It's soooo over-used and more of a catch-phrase than useful. What works for one situation may work for another. Many BPs are simply recommendations one can employ if they don't know what to do or they are a process fit to the situation of a company requirement.
Second, maybe this is a question better suited to the Tableau Server team. I find Tableau quite robust so it doesn't worry me. We're only using one server and 8 cores so what we have is probably "small" compared to others. Also, I'm really not in the mood to wake up between 12AM to 1AM, on a Sunday, to softly bring Server down.
Thinking about this, I guess one could make a small batch script that's on a schedule so it's triggered to execute at a certain time to stop the Server. Probably use another trigger to start it. I know Server starts up automatically when the virtual server is rebooted.
I would suggest to stop tableau services before the activity and turn back on once the IT has done with the upgrades. When you have a clustered environments it's always best to do a clean start/stop to make sure the workers are in Sync with Primary.
So my steps,
1) Send an email to all the stake holders about the downtime
2) Take a backup ( this way, you have the most recent copy)
3) Bring the tableau services down
4) IT does the upgrades
5) Restart tableau
6) Perform validation. ( Make sure all services are available, dashboards loading etc)
Vikram's steps above are how we operate (we're also on a distributed environment).
Would still be interesting to see what the Tableau Server team has to say though.
I agree with the steps mentioned by Vikram.
One thing I would like to emphasize is performing automated periodic( daily/once in 2 days) . This will ensure peace of mind .