Language Selection

English French German Italian Portuguese Spanish

Evaluating the performance of ext3 using write barriers and write caching

Filed under
Linux

The ext3 filesystem supports write barriers which are designed to allow a filesystem to take advantage of a disk’s write cache without fear of compromising the ingetrity of the filesystem on a power failure or kernel panic. I’ve seen reports that ext3 write barriers have a significant impact on performance.

Of course, the alternative is to disable the write cache (as is recommended many places) and suffer the performance loss from that. Given that the use of write barriers means forsaking many Linux storage tools which do not support write barriers (lvm, dmraid, dmcrypt, md raid except raid1, etc) it seems like the best option for flexibility is to just turn off the hard disks write cache and keep one’s options open. However, if the difference in performance between write barriers enabled and write cache disabled is significant then using write barriers would be a better option when IO performance is important. I often find IO to be the bottleneck on my servers so any improvement in IO performance is worth investigating.

I ran some tests with bonnie++ and came up with these numbers.




More in Tux Machines

CORD becomes a Linux Foundation project

Central Office Re-architected as a Data Center (CORD), an open source integrated solutions platform for service providers leveraging merchant silicon, white boxes, and open source platforms such as Open Network Operating System (ONOS), OpenStack, Docker, and the cloud operating system XOS, is now part of the Linux Foundation as a new independent project. The Linux foundation is already home to many open source networking projects, including OpenDaylight and ONOS, so CORD is a natural fit for the non-profit foundation. Read more

Google beefs Linux up kernel defenses in Android

Future versions of Android will be more resilient to exploits thanks to developers' efforts to integrate the latest Linux kernel defenses into the operating system. Android's security model relies heavily on the Linux kernel that sits at its core. As such, Android developers have always been interested in adding new security features that are intended to prevent potentially malicious code from reaching the kernel, which is the most privileged area of the operating system. Read more

Fork YOU! Sure, take the code. Then what?

There's an old adage in the open source world – if you don't like it, fork it. This advice, often given in a flippant manner, makes it seem like forking a piece of software is not a big deal. Indeed, forking a small project you find on GitHub is not a big deal. There's even a handy button to make it easy to fork it. Unlike many things in programming though, that interaction model, that simplicity of forking, does not scale. There is no button next to Debian that says Fork it! Thinking that all you need to do to make a project yours is to fork it is a fundamental misunderstanding of what large free/open source projects are – at their hearts, they are communities. One does not simply walk into Debian and fork it. One can, on the other hand, walk out of a project, bring all the other core developers along, and essentially leave the original an empty husk. This is what happened when LibreOffice forked away from the once-mighty OpenOffice; it's what happened when MariaDB split from MySQL; and it's what happened more recently when the core developers behind ownCloud left the company and forked the code to start their own project, Nextcloud. They also, thankfully, dropped the silly lowercase first letter thing. Nextcloud consists of the core developers who built ownCloud, but who were not, and, judging by the very public way this happened, had not been, in control of the direction of the product for some time. Read more

Proprietary and Microsoft Software