Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed last year.
Improve this question
I have a FreeBSD 12.1-RELEASE server and a CentOS 7 server. Both run on amd64.
I would like to set up a cluster file system, that runs on both platforms well. It should have CentOS 7 packages and FreeBSD packages. The solutions should be open-source software and "free of use".
After a little research, I found the following, but nontheless I always encountered drawbacks:
MooseFS3: Works on FreeBSD and CentOS, has packages for both, but only the MooseFS3 Pro version, which is commercial, has the functionality of real cluster functionality such as the possibility of mounting the file system from several nodes. Also I had locking problems with files that where access by my dovecot imap server daemon, when I run dovecot from the file system.
GlusterFS: Seems to work well, but there are no packages for the most current version of 8.x for FreeBSD. FreeBSD provides only a port for GlusterFS 3.x as of today. Different versions of GlusterFS can not operate together.
Ceph: Is very complex to configure, and I couldn't execute all of the steps of the official FreeBSD documentation for it, since the tool ceph-disk is deprecated in favor of ceph-volume. With Ceph-volume, though, I could not get it running with my zfs pool on FreeBSD, since the plugin for zfs for ceph-volume seemed to have some Linux code in it when it was ported to FreeBSD or similiar, so it might only run with ZFSOnLinux on Linux itsself.
OCFS2: I don't have much experience with that one, but its releases seem a bit outdated.
Lustre: No packages for FreeBSD and no acurate and up-to-date documentation how to set it up on a recent FreeBSD system
BeeGFS (Fraunhofer): No packages for FreeBSD, only for Linux
Hadoop MapR filesystem: Has a use case more for BigData storage than for a UNIX cluster filesystem, I don't know if it has FreeBSD packages.
So I don't find a good solution for a Cluster filesystem that runs on both FreeBSD and CentOS Linux. Even I'm planning to migrate the CentOS server to Fedora Server, so it should run there as well.
Anyone who can recommend me a recent compatible cluster file system that I could use on both FreeBSD and CentOS/Fedora Server and that allows real cluster file system features like replication and HA?
Or is there currently no cluster filesystem that fulfills my needs and I have to migrate the two machines running the same OS?
Thank you in advance.
Best regards,
rforberger
MooseFS3: Works on FreeBSD and CentOS, has packages for both, but only the MooseFS3 Pro version, which is commercial, has the functionality of real cluster functionality such as the possibility of mounting the file system from several nodes.
This is not true, you can mount MooseFS Community from as many nodes as you wish.
Glusterfs may be worth to try, it is based on fuse, which is available on FreeBSD, so you need only to build the userspace part, which may not be impossible, if it is not available for you OS version. On Linux it is definitely the simplest one to set up, since it comes packaged with most of the distros.
Lustre, despite supporting replicated directories, is more of a parallel filesystem oriented to HPC and high I/O performances, than a clustered filesystem oriented at redundancy, so I would not even consider it if redundancy is your purpose.
I have no experience with the other ones.
Related
Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
This question does not appear to be about a specific programming problem, a software algorithm, or software tools primarily used by programmers. If you believe the question would be on-topic on another Stack Exchange site, you can leave a comment to explain where the question may be able to be answered.
Closed 3 years ago.
Improve this question
PostgreSQL 11 works more efficiently in debian or redhat. Accordingly, I will choose a server and plan training. What are the advantages and disadvantages. Which operating system does the Postgres committee like?
There is no PostgreSQL committee. There is core, but they don't determine which operating systems are supported.
From the documentation (you probably read that):
A platform (that is, a CPU architecture and operating system combination) is considered supported by the PostgreSQL development community if the code contains provisions to work on that platform and it has recently been verified to build and pass its regression tests on that platform. Currently, most testing of platform compatibility is done automatically by test machines in the PostgreSQL Build Farm. If you are interested in using PostgreSQL on a platform that is not represented in the build farm, but on which the code works or can be made to work, you are strongly encouraged to set up a build farm member machine so that continued compatibility can be assured.
In general, PostgreSQL can be expected to work on these CPU architectures: x86, x86_64, IA64, PowerPC, PowerPC 64, S/390, S/390x, Sparc, Sparc 64, ARM, MIPS, MIPSEL, and PA-RISC. Code support exists for M68K, M32R, and VAX, but these architectures are not known to have been tested recently. It is often possible to build on an unsupported CPU type by configuring with --disable-spinlocks, but performance will be poor.
PostgreSQL can be expected to work on these operating systems: Linux (all recent distributions), Windows (Win2000 SP4 and later), FreeBSD, OpenBSD, NetBSD, macOS, AIX, HP/UX, and Solaris. Other Unix-like systems may also work but are not currently being tested. In most cases, all CPU architectures supported by a given operating system will work.
The majority of hackers use Linux, but there are people who develop on FreeBSD, MacOS or Windows.
If you know that PostgreSQL works most efficiently on certain Linux distributions, you know more than I do.
When choosing an operating system for PostgreSQL, I would proceed like this:
List the operating systems you are familiar with (or for which your organization has skilled administrators).
Exclude all operating systems for which there is not more than one animal in the buildfarm.
Exclude Windows.
Then pick any of these.
So this may seem unusual - but I am using a remote server with linux and gpus. The GPUS have old drivers and I do not have permission to update them. Can I do that within a virtualenv? The goal is then to use CUDA 10 +CUDNN and do all the nice things possible with the recent versions of these software.
virtualenv only concerns Python interpreter and its packages. It cannot handle hardware drivers - you absolutely need root permissions to install a hardware driver. Your best bet would be to contact the administrators of the server and ask them to upgrade. Alternately you can rent your own server, which is not at all as expensive as it used to be (e.g. from AWS).
I was asked to implement a High Availability and auto-failover in postgresql and I have been searching the internet all over to find the right architecture for that.
I have tried to do it in two architectures:
RedHat cluster
UCARP
Both ways came out pretty bad when I've meet bugs on UCARP restarting as master-master after several reboots and Redhat cluster fails to manage properly PostgreSQL service.
I would like to ask if anyone has ever succeeded to implement such an architecture and would like to explain me how or refer me to a tutorial that actually works.
The components that I have are: PostgreSQL at latest version (9.4), RHEL 6.6 and higher, all on virtual machines.
Thanks in advance,
Aviel Buskila.
I want to host the OSM (OpenStreetMap) locally. I need the basic idea what are required for hosting the OSM and how the task can be done in a step wise manner. I have to host it in Windows7 environment.
Any kind of help will be useful.
switch2osm contains detailed instructions and requirements for setting up a OSM server. If you have a Windows system then better set up a Linux VM inside it.
A bit too old but I will just put it here for someone who is searching for the same thing.
An exact instance of OpenStreetMap can be hosted locally by following the installation guide of OpenStreetMap.
Quoting from the Link:
"These instructions are designed for setting up The Rails Port for development and testing. If you want to deploy the software for your own project, then see the notes at the end.
You can install the software directly on your machine, which is the traditional and probably best-supported approach. However, there is an alternative which may be easier: Vagrant. This installs the software into a virtual machine, which makes it easier to get a consistent development environment and may avoid installation difficulties. For Vagrant instructions, see VAGRANT.md.
These instructions are based on Ubuntu 12.04 LTS, which is the platform used by the OSMF servers. The instructions also work, with only minor amendments, for all other current Ubuntu releases, Fedora and MacOSX
We don't recommend attempting to develop or deploy this software on Windows. If you need to use Windows, then try developing this software using Ubuntu in a virtual machine, or use Vagrant."
I'd like to ask for recommendations for a development environment with the following basic requirements:
on recent Fedora (e.g. 13 or 14) GNU/linux distribution
developing gtk+ apps with C
isolated (installation of files shouldn't affect the main system's system files)
allows for cheap snapshot of the entire development files
So far, the only two options I can think of are chroot jail and virtualized guest. I had a very hard time to get chroot jail to work properly with dbus (ended up snapshotting my entire system using btrfs). Also used recent VirtualBox which feels extra sluggish recently.
I would really like to know what the devs here use as their development environment.
I personally use two virtualized OpenBSD (CLI only) and Debian (w/ GUI) environments with backups, so I just wipe them clean when I need too. I use VirtualBox or VMWare.