Oracle Cloud – Your service is suspended due to exceeding resource quota !

Okay, had some good fun yesterday with my free trail for Oracle Cloud service for Database as a Service and Compute Cloud service. Even got connected to the environments via SQL Developer, although I didn’t get a SSH connection working (yet). I did my testing as planned and afterwards destroyed all environments. The latter just so to get new IP and other settings so people wouldn’t use the details given in the pictures yesterday to actually attempt to break in / hack the environments.

Me, being a bit bad boy myself, attempted to configure the 12cR1 High Performance environment as an In-Memory Column Store. This is officially “not allowed”, at least, thats what I figure reading the current differences between the High Performance and Extreme Performance offerings. But the In-Memory Column Store is easy to turn on (and off) as described in my posts;

• ORACLE DATABASE 12.1.0.2.0 – TURNING ON THE IN-MEMORY DATABASE OPTION
• ORACLE DATABASE 12.1.0.2.0 – TURNING OFF THE IN-MEMORY DATABASE OPTION

I didn’t finish my attempts to turn in on yesterday evening (hadn’t found the “start/stop/reboot” instance icon yet – as it turned of this morning), but anyway, “wax on, wax off”…   😎

Differences in High/Extreme Performance

In-Memory Column Store is not part of the High Performance offering. How can you look this up? It currently is mentioned on the “pricing” page of the main overview Database Cloud Service. In small fonts, it mentions…

High Performance

• Multitenant,
• Partitioning,
• Real Application Testing,
• Advanced Compression,
• Advanced Security,
• Label Security,
• Database Vault,
• OLAP,
• Advanced Analytics,
• Spatial and Graph,
• Diagnostics Pack,
• Tuning Pack,
• Database Lifecycle Management Pack,
• Data Masking & Subsetting Pack
• Cloud Management Pack for Oracle Database

Extreme Performance

• In-Memory Database,
• Active Data Guard,
• Multitenant,
• Partitioning,
• Real Application Testing,
• Advanced Compression,
• Advanced Security,
• Label Security,
• Database Vault,
• OLAP,
• Advanced Analytics,
• Spatial and Graph,
• Diagnostics Pack,
• Tuning Pack,
• Database Lifecycle Management Pack,
• Data Masking & Subsetting Pack
• Cloud Management Pack for Oracle Database

And as I saw yesterday while creating Extreme Performance environments, the ability under the “Extreme Performance” offering to create Real Application Cluster (RAC) environments, although there are some resource requirements to be met.

Your service is suspended due to exceeding resource quota

So yesterday, because I can (I thought), I happily created 4 new environments (after dropping everything beforehand – to be able to start with a clean slate) via clicking and entering all the details.

And when I left my terminal/browser yesterday, its showed the following configurations happily to be created:

• High Performance – 2 OCPU 11.2 database
• High Performance – 2 OCPU 12.1 database
• Extreme Performance – 2 node RAC 11.2 environment (using the 2 OCPU option)
• Extreme Performance – 2 node RAC 12.1 environment (using the 2 OCPU option)

This morning I had the following email and thought stuff like: “Oh yes, I am the only one who gets a free trail offer, and then is thrown out, after one day fiddling with it…”

(and other thoughts like “shouldn’t maybe have attempted to enable IMDB on HP”)

Preparing for the worse, I was relieved I was still able to login to the Identity Domain / Database Cloud portal pages.

Figuring out what’s wrong…

To my surprise, the four environments where happily running and had been created overnight. It even seemed (not tested) that I could access the environments. So whats wrong? What service bit was suspended / what was exceeding quota?

I clicked on the Dashboard button to get into the overview services page. Here I saw a clear message under the “Compute Cloud Service” section.

You will get a first real indication of what might be wrong, in the “Compute Cloud Service” section. In the meanwhile, this morning, I dropped the Oracle 11.2 RAC environment, thinking I overstepped by “trail” boundaries (but no clue yet why). The “Service Console” link will bring you to the “Compute Cloud Service” section mentioned in the post from yesterday. If you don’t know your quota restrictions out of your head, you will not get triggered by the (hint: overview details) information show there.

The “Details” link will redirect you to the “Service Detail” page for the “Compute Cloud Service”. You really have to look at the status info, because it doesn’t really pop out (scream?): “Suspended”. But okay, even if you notice, whats the reason, so you will be able to fix the problem. Up til now, the email didn’t give a clue nor the service Dashboard page.

So lets click “Metering” to see if there is some information that will shed some light on the “Why?”.

By the way you can create more graphs in this page and/or add additional info (CPU/network/etc) into the same graph if you wanted to.

Onwards to “Resource Quotas” then…

The following screenshot doesn’t reflect what I was initially seeing. As you might have noticed in the screenshot before, Block Storage Capacity (GB-Months) says “6.581“. The moment I created the next screenshot, my “drop the 11R2 RAC environment” action had just finished…so… Anyway. believe me when I say that the next screenshot reflects the “highlighting” of issues, that is “there was nothing in bold” or screaming at me otherwise: “You have only a BLOCK_STORAGE resource quota of  6000 and your using, see balance, 6581” (me: moron).

In all, my logic tells me, that that was the problem and when I, in my haste, dropped one of the two drop RAC environments (of was total 2 RAC environments, 2 single instance environments). Funny enough the other mentioned content / screenshots have not yet been updated with the updated figures. The details metering page still has the old value (and the dashboard page still shows an alert). So be aware.

Giving RAC another try…

So what would happen, if in the current situation with only 1 RAC environment and the 2 single instances, I would create exactly the same RAC environment which I just dropped. Would I now get a “Resource Limitation” alert or something, clearly stating that I should upgrade my service subscription?

So with a bit of “tongue in cheek” in mind, I created another RAC instance with the name “oneRACtoMany”.

If you follow the progress via the “In Progress” link of the Database service console pages, you will now see it failing due to lack of available resources provided via my service account.

In the “Instance create and delete history” section, if you click on it, it will show you all details from a certain defined time period (default is “last 24 hours”). When clicking on the “Details” link for the failed to create “onRACtoMany” it will state, in my case, that it failed due to a “Failed to create Compute resource for Database Server” reason.

In all, and I don’t know about you, but I would have had a bit more insight in the whole “Why and What?” information.

Even when failing directly, via my last attempt, it doesn’t signal that I am out of my “Block Storage Capacity”. Further more, my “click-click” exercise yesterday evening creating 2 RAC environments plus 2 Single instance environments, somehow succeeded. No warning at all, except the email message shown before.

Also in the overview “Instance create and delete history” of the Database Cloud service console, main page, all actions (except the “test create oneRACtoMany”) were successful. One issue on this page, b.t.w., which can become annoying, is the lack of presenting the messages ordering on “time” chronologically. And, there is no “knob” or method given of doing the chronological ordering ascending or descending.

Thinking about it

With this post I wanted to give you some additional insight about my test trails with the Oracle (Database/Compute) Cloud. I think there is some room for improvement when creating more useful web content. If you are, just like me, unknown what your resource limitations are, then it is not that clear where or what to look for when an problem was signaled via, for example, an email.

Even if you have found the “probable” cause, it is nether-less not 100% sure, that it actually is or was the cause. “Failed on lack of resources” can be caused by a lot of things. If not only from a customer perspective, the one thing you would want to know as quickly as possible, when your subscription is suspended due to a problem is, “what to fix”, “how to fix it” (and “how to add additional resource” & where to pay/etc). Apparently I succeeded yesterday evening in creating environments beyond my given subscription limits.

I think its in the interest of Oracle as well, to point as clearly as possible, where the problem is, how to fix it, so the signaled (production?) problem (eg. a stall) can be resolved.

But it starts with the ability to give a clear reason, identification or pointers to the : “What to fix?” (IMHO)

HTH/M.