![]() It is important for servers to have regular and stable updates, especially if they concern security and protection issues. All other things being equal, CentOS is more modest in resource consumption, so by a small margin, it can be considered a leader in the matter of stability. However, the development and testing of packages for CentOS are done by a professional company and they have a commercial interest in making sure everything runs stable. Both of them are good and are not complaining. Without being too crude, it is impossible to single out any OS from these two in terms of stability. For most tasks, it will be more than enough. The latter has much fewer packages, but still has everything you need to set up and prepare your server for work.įor CentOS, the preferred and most convenient repository is the YUM repository with the. Debian easily beats even Ubuntu in the number of repositories, not to mention the rather conservative CentOS. And if you need any other third-party PPA repositories, you can hook them up in no time at all if you need highly specialized software. Where are the best repositories: Debian vs CentOSĭebian is the absolute leader here - it has the most embedded packages. Another advantage of CentOS compared to Debian is the stricter quality control - since the development team is stable and hardly ever changes, it has a positive effect on the quality of all new releases and updates. This OS is very good because of its high stability and conciseness, it has a long support cycle - up to 10 years - which makes it a good choice for companies looking for a stable server solution for the long term. The disadvantages are the short support time - on average each release of Debian is supported for about 3 years.ĬentOS, on the other hand, is based entirely on Red Hat Enterprise Linux, but it is distributed for free. Basically, the community does a very good job with the support and no questions asked about the latest releases, everything works fine and stable but some of the problems are not fast enough and important updates can take weeks to get approved. It is a steadily evolving OS, created by modifying a pure kernel. To begin with, Debian is supported in its entirety by volunteers, at no cost to you. Where are the best repositories: Debian vs CentOS.CentOS 7.4 performance without SW RAID is only slightly better than FreeBSD 11.1 with ZFS RAID (for TCP-B and 100 concurrent clients).What you will learn about in this article? The results show the inefficiency of the Linux SW RAID (or ZFS RAID efficiency). I assumed that such a difference was caused by the Linux software RAID overhead, so I did three more TCP-B benchmarks for 100 concurrent clients, this time without software RAID: I was positively surprised by FreeBSD 11.1 which was more than twice as fast as the best performing Linux, despite the fact that FreeBSD used ZFS which is a copy-on-write file system. Among the GNU/Linux distributions, Centos 7.4 was the best performer, while Debian 9.2 was slowest. Unfortunately I did not find out what caused such a mediocre FreeBSD performance.Ī more realistic image of PostgreSQL performance were provided by read write (TCP-B) benchmark. The best peforming OS in read only benchmark was openSUSE 42.3, while FreeBSD was about 40% slower. The benchmark had shown that the difference between various GNU/Linux distributions is not very significant. During the test, I gradually increased the number of cuncurrent clients accessing the database. Before each read write benchmark, pgbench tables were vacuumed (the -v flag was used). The test script consisted of three parts: vacuum + warmup, read only benchmark and read write benchmark. The pgbench program version 10.1 running on a separate FreeBSD 11.1 machine was used to generate the load. 200GiB TCP-B: read write, the dataset fits into the PostgreSQL cache.32GiB TCP-B: read write, the dataset does not fit into the PostgreSQL cache.200GiB read only: read only test, the dataset fits into the PostgreSQL cache.32GiB read only: read only test (only selects, without data modifications), the dataset does not fit into the PostgreSQL cache.I was interested in these benchmark combinations: ( ): Intel Xeon E5-1650 v3, 6 cores, 12 threads, 256GB DDR4 ECC RAM, 2 x 480 GB SATA 6 Gb/s Data Center Series SSD, was used as PostgreSQL server. ![]() ( ): Intel i7-6700, 4 cores, 8 threads, 32GB DDR4 RAM, was used for generating SQL queries using pgbench.The test infrastructure consisted of two dedicated servers connected with a 1 Gbit/s network:
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |