Ubuntu 14.04 LTS vs. Red Hat Enterprise Linux 7 Benchmarks
Phoronix: Ubuntu 14.04 LTS vs. Red Hat Enterprise Linux 7 Benchmarks
Last week when Red Hat Enterprise Linux 7 Beta 1 was released I was already running RHEL7 benchmarks looking at the performance of Red Hat Enterprise Linux 7 over RHEL 6.5. In this article for some extra benchmarks to put out over the weekend is a quick comparison of Ubuntu 14.04 LTS in its current development state against Red Hat Enterprise Linux 7 Beta 1.
Ubuntu is better looking, productive, faster and FREE!
RHEL7 better only for professional use.
Don't forget that RHEL has SELinux enabled by default and Ubuntu doesn't.
I'd like see results with SELinux turned on on Ubuntu and/or SELinux turned off in RHEL (I prefer the first).
Personally I would never run a Linux that has SELinux turned off....
Excuse my oblivion, doesn't it have AppArmor activated by default?
Originally Posted by fhuberts
Ubuntu LTS doesn't even use a long term kernel version, i think Debian 7 and the upcoming CentOS and Scientific Linux 7 are still far more apropriated for servers
Originally Posted by phoronix
I think testing like that would be very appropriate for benchmarking long-support enterprise-targeted OS releases. With an OS targeted at residential users, sticking to the out-of-the-box defaults makes the most sense as I agree with the adage that the common operator is unlikely to change such defaults. However, long-term enterprise releases are often intended to be deployed in quantity (both on server and desktop), and system administrators are much more likely to tweak the configuration for their environment's particular needs. Therefore, I think it makes sense to try to establish some common base for benchmarks that is not influenced by functionality that is likely to be tweaked anyway (which I think SELinux and AppArmor falls under).
Originally Posted by fhuberts
As for the default filesystems, I don't really have any requests or suggestions other than maybe to say that there's no need to even bother at all. If a sysadmin wants to see how XFS impacts RHEL's performance vs what it would have been under Ext4 or something else, that sysadmin is better off looking at some of the dedicated FS benchmarks that have already published by Michael.
One thing I have always wished to see in comparisons like this one, is NFS performance. In my ~16 years of work, its all been on UNIX/Linux. For the most part the local disk in the server or workstation is only used to hold the OS. All data and apps are ran off NFS mounts. Back about 5 years ago, we went from RHEL3 -> RHEL5 on all machines, we noticed very decent read/write performance increase, with same backend storage(NetAPP)
It we be cool to see the differences between RHEL 6,7, Ubuntu on how they do the NFS. Here's an idea of a test setup
NFS Server: Just setup any Linux OS and export out a volume. NFSv3?
Networking: Nothing fancy, just GigE switch, all machines the same.
Test Machines: Use default mount options the OS wants to use. Could play around with this a bit as different benchmarks
As for the test, just use the same IO benchmarks you would, but running against the NFS mount. The test app doesn't need to run there, just its read/write stuff would be on the NFS.
Anyone else ever want this, are there articles on this topic that I have missed before?