SoylentNews is people
SoylentNews
from the support-costs-versus-business-output dept.
My organization recently deployed about 60 servers running Oracle Linux. I have worn a few different hats (SW Engineer, Systems Administration, Help Desk) over about 12 years of computing experience, and I have only ever had to use tech support for an Linux OS related problem once, and it was related to converting Red Hat classic subscriptions to the new subscription manager. I have developed software using C, C++, and Python. I have also resolved problems with open source software by downloading the source, fixing bugs, recompiling, and finally submitting the fix upstream.
Other than having someone to yell at, is there any benefit to paying for support? Would be better to just set aside some cash to pay a consultant if there is something I can't figure out myself?
Original Submission
(Score: 1) by VanderDecken on Wednesday June 03 2015, @03:05PM
I know the OP was asking specifically about Oracle Linux, however I'd like to address the question in the wider context of FOSS as well.
Even when you do all of your troubleshooting on your own, having at least a few (or one) system under paid support will often provide you access to customer-only knowledge databases. I've been in this situation where a client mostly has CentOS systems, but I've had them buy a few RedHat licences to give them such access. Plus, the RedHat licenses were then used in test environments to test/validate software on upcoming OS versions, and ensuring that new hardware products were going to be suitable for future production systems.
For Oracle Linux, you should also consider what else is running on the box. For example, is it running Oracle Database? If so, speak to your DBAs: I'm sure that they will have firm opinions (one way or another) as to whether support is needed for Oracle Database. If it is, you almost certainly need support for the underlying Oracle Linux. You don't want to be taking a database outage, trying to open an emergency support ticket with a paid DB subscription only to be told, "sorry, we won't help you because your OS isn't under support". And don't forget with Oracle in particular, you don't want to be in a situation where you are *supposed* to have a license but do not. They are unforgiving during their licensing audits.
Another argument when it comes to FOSS is that even if you don't need the support, it is in your best interest to support those who provide the products on which you depend. This (in pragmatic terms) is especially true if the product is being provided by a smaller community. For example, I ensure that I make donations to OpenBSD every year -- this should be more widespread when you consider all the commercial and FOSS systems that benefit from the core infrastructure software that is maintained by the OpenBSD group. If giving more money to Oracle sticks in your craw, then consider other options like the Linux Foundation. As individuals we should certainly be supporting FOSS this way. As corporations, we should not only be supporting them but are more capable of doing so in a significant manner.
Next, being able to say to management and/or auditors, "yes, we have support contracts in place for these systems" can be valuable; often being able to say that you *can* have paid support may be a determining factor for management as to whether they allow you to deploy the system. Of course, some organizations care only about reducing the annual licensing fees, so it depends on the circumstances.
I would suggest evaluating the support question on an environment-by-environment basis. ie: The answer may be different for DB servers vs application servers vs web servers vs DNS servers, etc. How about production vs test vs development machines? In addition to the considerations above, if these systems are currently Oracle Linux and you stop licensing them, do you also stop getting security updates? How much is it going to cost you in time and money to migrate to a comparable system where you *will* get necessary updates? (CentOS, Scientific Linux, FreeBSD, etc).
The two most common elements in the universe are hydrogen and stupidity.