Re: Usage of network file systems with vserver
Sandino Araico Sánchez <[email protected]> Sat, 18 Nov 2017 06:11:42 -0600
| Newsgroups | gmane.linux.vserver |
|---|---|
| Message-ID | <[email protected]> |
On 18/11/17 03:13, Oliver Welter wrote: > Hi All, > > I know that this is a bit offtopic but I hope to get some > pointers/experiences... > > I am currently using a dual-node setup with DRBD to have redundancy > for my vserver guests. This was a nice setup when I started with two > nodes, but as I am now running 3 pairs which need some hardware > refresh I am looking for alternative solutions to share my guest data > over multiple nodes. > > My current setup is to mount the block device from the guest fstab so > the data is only visible inside the guest. This was a major problem > for me when I tried using GlusterFS some years ago. > > As my budget does not cover an invest in a "real" SAN/FibreChannel > backend, I am looking for practical experience with things like > GlusterFS, Ceph or other solutions based on decent "standard servers". > Are those systems cooperating with the vserver approach and are there > any things to take special care of? > > best regards > > Oliver > Shared network filesystems I have them mounted on the host filesystem. This happens before util-vserver starts and you use the _netdev option in the hosts's fstab to ensure the filesystem is mounted after network . In the guest's config fstab I use -o bind to have the already mounted shared filesystems bind-mounted on the guest none /proc proc defaults 0 0 none /dev/pts devpts gid=5,mode=620 0 0 none /dev/shm tmpfs defaults 0 0 /shared/home /home none bind,rw 0 0 /shared/vhosts.d /etc/apache2/vhosts.d none bind 0 0 -- Sandino Araico Sánchez http://sandino.net