• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

How do you share out your data?

R

Red Squirrel

[H]F Junkie
Joined
Nov 29, 2009
Messages
9,217
Currently I use NFS, but I find it can be so flaky. One of my servers randomly hiccuped and wont recover, all the NFS shares are locked up, can't unmount, can't do anything, if I try to do anything like cd or dir the folder, it just locks right up. It's a mess.

I also hate the fact that when I make changes I have to restart the service, then it knocks everything offline. There's got to be something better than NFS.

iSCSI wont really work for my needs since it's technically only meant for one host at a time. I have lot of shares that are on multiiple machines.

I use this for everything from regular files, to VMs. Basically, I want it to act as close as possible to like a local disk, but it still needs to be file aware and allow sharing across multiple systems, and not a block device.

Is there something out there that's better than NFS? For Linux obviously.
 
Mount your exports "soft". The client will then not hang but instead filesystem access will fail. Pick your poison. You could also fix the hiccups instead. A normal NFS mount will recover when the server is alive again.

With changes I suppose you mean changes in /etc/exports? You only need to restart mountd for that. This doesn't affect existing mounts.

Granted, NFS is a pile of dogshit, but most of your problems seem to come from doing it wrong.

Edit: I'm assuming you don't know this, but NFS is not a single daemon, it's a conglomerate of services revolving around the RPC protocol. You have the rpcbind daemon managing the mapping of "RPC services" like mountd, nfsd, statd, lockd to TCP and UDP port numbers. Well, and then you have mountd managing client access according to /etc/exports and nfsd managing the actual filesystem access of clients. statd and lockd are additional daemons than can handle file locking and stuff. It's all pretty arcane.

Look around on your system with man -k rpc, man -k nfs and so on.

Code: 
# rpcinfo -p localhost
   program vers proto   port
    100000    2   tcp    111  portmapper
    100000    2   udp    111  portmapper
    100005    1   udp    705  mountd
    100005    3   udp    705  mountd
    100005    1   tcp    902  mountd
    100005    3   tcp    902  mountd
    100003    2   udp   2049  nfs
    100003    3   udp   2049  nfs
    100003    2   tcp   2049  nfs
    100003    3   tcp   2049  nfs
    100021    0   udp    665  nlockmgr
    100021    1   udp    665  nlockmgr
    100021    3   udp    665  nlockmgr
    100021    4   udp    665  nlockmgr
    100021    1   tcp    797  nlockmgr
    100021    3   tcp    797  nlockmgr
    100021    4   tcp    797  nlockmgr
    100024    1   udp    659  status
    100024    1   tcp    897  status
    100026    1   udp    743  bootparam
 
Last edited:
You must log in or register to reply here.
Back
Top