You don’t need or want lamp.
With digital ocean (or any server) you’ll want to do some basic things like install fail2ban and enable automatic os upgrades. No, it is not managed.
You don’t need or want lamp.
With digital ocean (or any server) you’ll want to do some basic things like install fail2ban and enable automatic os upgrades. No, it is not managed.
Thank you!
In Home Region (Your home region contains your account information and identity resources. It is not changeable after your tenancy is provisioned. If you are unsure which region to select as your home region, contact your sales representative before you create your account: Cloud Data Center Regions and Locations | Oracle) what should I choose?
Thanks
EDIT:
I registered and I’m starting to install
I have installed Discourse on Oracle Cloud!!!
Now, I would like import Mybb database (Importing MyBB to Discourse), but apt doesn’t work
[root@oecdiscourse discourse]# sudo apt update
sudo: apt: command not found
[root@oecdiscourse discourse]# apt-get update
bash: apt-get: command not found
[root@oecdiscourse discourse]# sudo apt-get update
sudo: apt-get: command not found
[root@oecdiscourse discourse]# cd
[root@oecdiscourse ~]# sudo apt-get update
sudo: apt-get: command not found
Is it installed on the server?
You probably installed their version of Linux rather than Ubuntu. You’ll need to use Ubuntu or rtfm to install stuff with their os.
The Oracle Linux is more Redhat than Ubuntu. Try yum
versions of those commands.
Does anybody know how this would perform in comparison to the smallest droplet at Digital Ocean? I can’t figure out how to actually compare them.
The current cheapest droplet appears to be $5/month for 1 shared vCPU, 1 GB ram, 25 GB SSD (no swap partition), 1000GB transfer. The free Oracle host is 1 OCPU (defined as “CPU capacity equivalent of one physical core of an Intel Xeon processor with hyper threading enabled”), 1 GB ram, 50 GB SSD (10 as swap), and I think 1000GB transfer.
I have not yet figured out how to see how much data transfer I’m using.
Hmmm. Elsewhere I see Oracle as offering 1/8th OCPU. Details of the Always Free Resources But you can have two of them, so you can split DB and front end if desired.
Is there a way to compare the cpu capability? Maybe a command that could be run on each? Seems like Oracle has a slight edge, but that’s ignoring the CPU power, which is a black box to me.
Easy way is to run a rebuild of standard Discourse in each and time how long it takes.
I get about 20 minutes rebuild time, maybe someone with a small droplet can offer a comparison. From time sudo /var/discourse/launcher rebuild app
leaving out the upgrade Postgres rebuild:
Apr 17 2020: 22m16.598s
Jun 2 2020: 18m28.863s
Jun 28 2020: 20m27.314s
I’ll run it on a D.O. small droplet. But how do I leave out a Postgres rebuild?
I meant I left out the one that had Postgres 10 to 12 update. That took longer, but was doing more.
There are plenty of variables about how old the Discourse image is and how much work it has to do to update itself – but a really good rebuild time is more like 5 minutes. And a decent rebuild time would be 10 minutes.
That makes sense. Mine just finished in 6.5 minutes. But I had updated it earlier today… I wonder, @elijah what the second rebuild would take on your site the same day (ie similar to mine).
I know how to pay money for faster service. But it is nice to see the comparison. Maybe I’ll split db to a separate host at some point and redo the test.
@jtbayly: I didn’t save exact times, but early on while fiddling with what plugins to use I did some back to back rebuilds. There was not much difference in time.
Hmmm… I guess @codinghorror or others could shed more light on that, but I doubt it would have much impact on the rebuild speed. It doesn’t seem to me like a lot of the rebuild is spent in the DB.
So a free Oracle server is ~4x slower than the cheapest DO droplet and ~7x than a High CPU DO droplet (rebuild takes 3 minutes)
Performance tip: block volumes performance is dependant on disk size (60 IOPS/GB and 480 KB/s/GB). If you’re using only one free vm you can allocate your 100GB of free storage to that vm to double the IO. It’ll speed rebuild times quite a bit.
A part of the definition is missing: Each OCPU corresponds to two hardware execution
threads, known as vCPUs. That is, an OCPU is equivalent to two vCPU in other clouds (like DigitalOcean). Although of course in the case of the free service you don’t even get a full OCPU…
And a heads up of the Oracle TOS in case you hadn’t read it:
1.3 You may not, and may not cause or permit others to: … (b) perform
or disclose any benchmarking or availability testing of the Services;
You can rent a server billed hourly at a hosting company with this kind of billing and do the tests yourself. It will take some time to install and build Discourse on the servers, but you’ll have all the specs you want and it will cost you a few cents to a few dollars.
A TOS that prevents you from testing whether your website is staying online? Whether they are providing what you are paying them to provide. Whether your server is functioning. Whether people can buy your products. In fact, it prohibits having a website on your server at all, since that permits others to perform an availability test of the Service.
So idiotic. If people actually followed the “rules” that lawyers write for everybody, the world would come to a standstill.