Solution of Mount timeout problem for Linux old NFS system

Background Introduction

Because of the business needs, the school to transplant the test system to the cloud platform, the first step is to first put the old system a NFS storage platform mount to the new system B. The old test system A has been in service for a long time, the operating system is still the REDHAT4 series, the version is Linux 2.4.20.

problem

At the end of the B mount, there is a problem with the input

Mount-  t NFS 10.77.30.31:/opt/oj/contests  /MNT/NFS

Super times wrong

A-side NFS service, other older systems can mounta NFS files, but Linux 3.0 after the server is still not mount. Later asked the others, the original they have already known, but has not been resolved.

Tracking

First on the B-terminal showmount look, found everything normal

[Email protected]:/mnt# showmount-e 10.77.30.31Export list for 10.77.30.31:/opt/oj/contests (everyone)

This indicates that the NFS service on the B-side is operational and the firewall is not blocked.

Then find the log on the A-side, and find that the error has been circulated

Nfsd:unexporting all Filesystemsnfsd:last server have exitednfsd:unexporting all filesystems

So online Baidu This problem, are said to be NFS server /PROC/FS/NFSD not loaded, so looked under, found that there is no such thing as a end

So try mount, find no, prompt error

Mount-t nfsd  nfsd  /proc/fs/nfsd mount:mount POINT/PROC/FS/NFSD does not exist

This type of mount on the B-terminal or other newer systems is normal

Then a-end is not installed NFSD service, although the basic impossibility, but still verify that a Terminal NFSD service module is loaded,/proc also has

[Email protected] contests]# lsmod|grep nfsdnfsd                   81104   8  (AutoClean) lockd 59536   1  ( AutoClean) [NFS Nfsd]sunrpc                 87516   1  (AutoClean) [NFS nfsd lockd][[email protected] contests]# ls/proc/ Net/rpc/nfsd/proc/net/rpc/nfsd

solution to the problem

By this time I have confirmed that the basic a-side NFS service version is too old for the problem. But since A's operating system is too old to install a new version of NFS, then there is no way.

I searched for "nfsd:unexporting all filesystems" in the course of the problem, found two links http://blog.csdn.net/zanget/article/details/ 6659314,http://www.233.com/linux/instructs/060210/172416755-2.html mentioned nfsvers=x, which I instinctively felt the problem. Further reading has been found to be used to select the NFS version for optimization. Through the Nfsstat, the end of a only support 1-3 NFS version, and now the default is 4, so on the B-side tried the next parameter, found success.

[Email protected]:/mnt# mount-  t NFS 10.77.30.31:/opt/oj/contests  /mnt/nfs-o nfsvers=3[email protected]:/mnt # Ls/mnt/nfsproblems  problems.1  runs  work  working

Summary

This kind of problem first to find the log information of the problem, but can not rely on the online universal solution, but should understand its mechanism. Finally thanks to the next two authors of the cited article:)

Solution of Mount timeout problem for Linux old NFS system