0

I'm looking into to using GlusterFS and am struggling with the GlusterFS_Concepts explanation here.

Disadvantages - If you lose a single server, you lose access to all the files that are hosted on that server. This is why distribute is typically graphed to the replicate translator.

GlusterFS is a replicated, distributed file system, why would the loss of one server cause you to lose access to all files on that server? I feel I am missing something here. Surely one of the main points of replication is that I would be able to access the files even in the event of a server failure?

user2644013
  • 192
  • 2
  • 14

2 Answers2

1

If you lose a single server, you lose access to all the files that are hosted >on that server. This is why distribute is typically graphed to the replicate >translator.

This refers to distributed volumes that do not have replication built in. A replicated, distributed-replicated or erasure coded volume in Gluster can handle single server failures gracefully.

Vijay
  • 36
  • 1
0

Replication provides for redundancy. However, some gluster installations are all about enormous filesystems (distributed) and not about redundancy (replicated).

Imagine your process creates vast (enormous) temporary data sets. You might well have no need for replication.

user3718260
  • 181
  • 1
  • 6