5 ms·

NFS in a modern system seems like a huge red flag.

existencebox 10y ago
Speaking as someone who has maintained some rather large NFS utilizing systems/datacenters in the last decade, and found it perfectly suitable to my needs, I'd be curious to why you'd say that?
- brobinson 10y ago
  Not a block device, so not having cgroups (read/write bps limiting + iops limiting) is one annoyance of NFS in large deployments
  - Karunamon 10y ago
    I'd also add extreme fault intolerance. Even when you're mounting soft, or hard+intr, NFS failures have a way of requiring reboots to fully fix.
    - Scaevolus 10y ago
      The intr mount option has been a no-op for 8 years. /bin/umount does an fstat before the umount syscall, so if an NFS mount is broken, it will probably hang forever instead of unmounting it! You can invoke the syscall directly to get the behavior you need.
      - acdha 10y ago
        That hasn't been dependable in my experience - the best workaround I've found is either a lazy unmount + remount if you can tolerate the hung process taking the full NFS retry timeout to unblock or downing the network interface to accelerate the process.
    - xir78 10y ago
      This is fair to a degree, but vendors have VIPs that make failures fairly transparent. But if you need low latency on a large data set is there something better? Where low latency means <100ma reads
  - X86BSD 10y ago
    So use Jails, ZFS and rctl. Problem solved.
    - jamespo 10y ago
      How does that fix a client accessing network storage?
      - X86BSD 10y ago
        You can use Samba with it. Unless you absolutely require NFS, samba is far more robust and less prone to needing to be hard rebooted.
- xorgar831 10y ago
  I'd agree, it's hard to beat if you need high bandwidth and low latency access to large data sets.
- beachstartup 10y ago
  throwing away 30+ years of protocol development for the latest and greatest containerization paradigm is what the cool kids south of market street talk about at happy hour.
cosmie 10y ago
What's wrong with NFS? If you need a shared, POSIX-compliant drive between multiple hosts, there's not really an alternative. Even the fairly new AWS Elastic File System service is just a NFS4.1-based managed service
- jacques_chester 10y ago
  While the grandparent comment might be a little too emphatic, surrendering the flexibility of file stores in favour of object stores does lend itself to permitting nice non-functional properties.
- falsedan 10y ago
  You can pay money for a SAN, which is $$$$ & enterprise but is a viable alternative…
  - jo909 10y ago
    A SAN gives you block storage - virtual hard disks. You will need a cluster filesystem to be able to use the same virtual disks on multiple machines concurrently, which are not easy to operate and have their own bunch of performance problems.
- fh973 10y ago
  Some of the file systems come with native clients. I'd say for container environments that's a viable alternative to NFS. When done properly the native client can do low latency as it usually has the minimum number of network hops and do failover without requiring things like virtual IPs.
superuser2 10y ago
If you want volumes that can be dynamically attached to any machine, you're not buying a SAN, and you're not in a cloud provider with a portable block storage abstraction, what do you propose? About the only thing I can think of is Ceph, and it stands to reason there is a lot more NFS than Ceph expertise on the job market.

Speaking as someone who has maintained some rather large NFS utilizing systems/datacenters in the last decade, and found it perfectly suitable to my needs, I'd be curious to why you'd say that?

brobinson 10y ago
Not a block device, so not having cgroups (read/write bps limiting + iops limiting) is one annoyance of NFS in large deployments
- Karunamon 10y ago
  I'd also add extreme fault intolerance. Even when you're mounting soft, or hard+intr, NFS failures have a way of requiring reboots to fully fix.
  - Scaevolus 10y ago
    The intr mount option has been a no-op for 8 years. /bin/umount does an fstat before the umount syscall, so if an NFS mount is broken, it will probably hang forever instead of unmounting it! You can invoke the syscall directly to get the behavior you need.
    - acdha 10y ago
      That hasn't been dependable in my experience - the best workaround I've found is either a lazy unmount + remount if you can tolerate the hung process taking the full NFS retry timeout to unblock or downing the network interface to accelerate the process.
  - xir78 10y ago
    This is fair to a degree, but vendors have VIPs that make failures fairly transparent. But if you need low latency on a large data set is there something better? Where low latency means <100ma reads
- X86BSD 10y ago
  So use Jails, ZFS and rctl. Problem solved.
  - jamespo 10y ago
    How does that fix a client accessing network storage?
    - X86BSD 10y ago
      You can use Samba with it. Unless you absolutely require NFS, samba is far more robust and less prone to needing to be hard rebooted.
xorgar831 10y ago
I'd agree, it's hard to beat if you need high bandwidth and low latency access to large data sets.
beachstartup 10y ago
throwing away 30+ years of protocol development for the latest and greatest containerization paradigm is what the cool kids south of market street talk about at happy hour.

cosmie 10y ago

What's wrong with NFS? If you need a shared, POSIX-compliant drive between multiple hosts, there's not really an alternative. Even the fairly new AWS Elastic File System service is just a NFS4.1-based managed service

jacques_chester 10y ago
While the grandparent comment might be a little too emphatic, surrendering the flexibility of file stores in favour of object stores does lend itself to permitting nice non-functional properties.
falsedan 10y ago
You can pay money for a SAN, which is $$$$ & enterprise but is a viable alternative…
- jo909 10y ago
  A SAN gives you block storage - virtual hard disks. You will need a cluster filesystem to be able to use the same virtual disks on multiple machines concurrently, which are not easy to operate and have their own bunch of performance problems.
fh973 10y ago
Some of the file systems come with native clients. I'd say for container environments that's a viable alternative to NFS. When done properly the native client can do low latency as it usually has the minimum number of network hops and do failover without requiring things like virtual IPs.

superuser2 10y ago

If you want volumes that can be dynamically attached to any machine, you're not buying a SAN, and you're not in a cloud provider with a portable block storage abstraction, what do you propose? About the only thing I can think of is Ceph, and it stands to reason there is a lot more NFS than Ceph expertise on the job market.