I needed a good NFS storage cloud in AWS, and since EFS is just in preview mode, I did some research, and GFS looked like it would be the best bet.
Currently, I have two servers that are part of one cluster, each is using an EBS volume for the storage, and I was able to create them just fine, and I can read/write as well. But I think I'm having speed issues.
One of the volumes is shared between two web servers, and it holds the php files, which uses the Laravel framework, so theres a lot of files to install. Checking out the git repo and installing via npm install basically never ends, but if I go to a separate directory (not on NFS), and check out the same repo, it works fine, then I just rsync everything over (which again, takes much longer than I would expect, hours). The directory size of everything checked out is only 255M, so I wouldn't think this would take that long.
I created a 1GB file using dd to see the speed:
# dd if=/dev/zero of=./test.txt bs=1024k count=1000 conv=sync
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 139.762 s, 7.5 MB/s
7.5 MB/s - Thats pretty slow I would think.
Would something like this be improved by adding more resources to the Glusterfs nodes? or adding more nodes to the cluster? Or is there throttling somewhere? I'm somewhat new to AWS, as well as GlusterFS, so any help would be appreciated.
UPDATE
Just an FYI, both instances are t2.micro, and in the us-west-2a AZ
0 Answers