Support for changing the CPU and/or Memory Count for Compute Flex shapes has been added to the OCI-AutoScale Script.
By specifying a (cpu:memory) value in the schedule, the compute instances will be update accordingly.
IMPORTANT: This will reboot the instance!
Script seems very promising but unfortunately running into a problem. It works great to take an instance down per the schedule for some reason for me (obviously not a widespread issue) it isn’t bringing a compute instance back online for a non-zero hour entry. Any idea what might be causing this? I’d appreciate any help/thoughts you can provide on this. I’m glad it’s working well for many others and help I can so leverage it too 🙂
# Running Down actions
## python3 /home/ubuntu/OCI-AutoScale/AutoScaleALL.py -a Down -ip -ignrtime -printocid
Starting Auto Scaling script on region us-gov-ashburn-1, executing Down actions
Ignoring Region Datetime, Using local time
Day of week: Thursday, Nth day in Month: 3, IsWeekday: True, Current hour: 21, Current DayOfMonth: 19
Getting all resources supported by the search function…
Finding MySQL instances in 2 Compartments…
ee
Checking 1 Resources for Auto Scale…
Checking dsop006_test (Instance) – ocid1.instance.oc3.us-gov-ashburn-REDACTED
– Active schedule for dsop006_test: 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,2,0,2,2,[0],2,0,
– Initiate Compute VM shutdown for dsop006_test
Waiting for all threads to complete…
Region us-gov-ashburn-1 Completed.
All scaling tasks done, checked 1 resources.
# Running Up actions (no message showing bring up and confirmed in OCI Console)
## /home/ubuntu/OCI-AutoScale/AutoScaleALL.py -a Up -ip -ignrtime -printocid
Starting Auto Scaling script on region us-gov-ashburn-1, executing Up actions
Ignoring Region Datetime, Using local time
Day of week: Thursday, Nth day in Month: 3, IsWeekday: True, Current hour: 22, Current DayOfMonth: 19
Getting all resources supported by the search function…
Finding MySQL instances in 2 Compartments…
ee
Checking 1 Resources for Auto Scale…
Checking dsop006_test (Instance) – ocid1.instance.oc3.us-gov-ashburn-REDACTED
– Active schedule for dsop006_test: 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,2,0,2,2,0,[2],0,
Waiting for all threads to complete…
Region us-gov-ashburn-1 Completed.
All scaling tasks done, checked 1 resources.
what size is the VM in amount of OCPUs? If it is a none flex shape, you should only use the value 0 for off and 1 for on (independent of the size)
Thanks. The compute instance is using an AMD flex shape (VM.Standard.E3.Flex) but for some reason it doesn’t treat it as such. Changing to use 1/on brought the instance up right on schedule and did not reduce OCPU count which is exactly what I needed. I really appreciate your scaling script and other articles you’ve done.
Also, an article topic you might consider at some point is one on the OCI CLI helper script “Project O” — I really like O but would be surprised if I didn’t learn some new tricks from such an article.
It has 2 OCPUs.
It seems for what I was trying to do should always be using just 0, 1 or * regardless of actual CPU count. I haven’t tried it but I’d bet had I used the ‘(cpu:ram)’ notation it would have done a reboot/resize. From some of the explanatory content though it seemed it was necessary to use the desired (if CPU higher >1 count) for CPUs in the schedules but that doesn’t seem to be the case — at least for me. Really great script. Super helpful!!
For the tag value, does the first position indicate the midnight hour? or the 1:00 AM hour?
It is 0 hour, so between 0:00 and 0:59
Hello!, VERY nice litle script you have done 🙂
I have got the most functions to work, except one thing. It cant find our AutonomousTP Database on dedicated Exadata. Is it supposed to find that? i see in the logs that if finds all other scheduels. except that one.
Best Regards Martin
No I do not think the script support that. I would need to look into that, if I can test that in my own account. I assume it does allow power on/off and scaling thru the OCI Console?