Sponsored Content
Top Forums UNIX for Advanced & Expert Users NFS active/active cluster with gluster and NFS-ganesha, I can create dirs, but not files! Post 303045800 by Linusolaradm1 on Monday 13th of April 2020 05:52:02 PM
Old 04-13-2020
NFS active/active cluster with gluster and NFS-ganesha, I can create dirs, but not files!

Following this fantastic guide I finally setup an active working active/active nfs4 cluster.
The setup is not too difficult: two servers with ganesha+glusterfs other two with haproxy and keepalived.
The cluster works fine, from one client I have mount the nfs share called "nfshare" without problem


Code:
mount -t nfs4

nfsserver:/nfsshare on /mnt type nfs4 (rw,relatime,vers=4.2,rsize=1048576,wsize=1048576,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,clientaddr=10.2.0.170,local_lock=none,addr=10.2.0.7)


I have created some dirs on mnt and works perfect(there are replicated on nfsgluster servers). I have also rebooted for testing the nfs server one or two, and after a very little downtime (2 seconds) the connection is up and working.
The problem is..mkdir works, but every file operation(dd,cp,touch) fails(stuck,freeze until ctrl+c is pressed).

On both the nfs servers and keepalived servers there is no problem with selinux


Code:
grep -i denied /var/log/audit/audit.log


What can be?Why mkdir works and file operation not?

--- Post updated at 05:52 PM ---

Solution found, ganesha was misconfigured, I have forgot to add the ports for services.
 

We Also Found This Discussion For You

1. AIX

Very Active NFS and Disk Usage

Here is the scenario... NFS share that is accessed every few minutes by approx 70 systems (AIX 5.3/6.1). Filesystem space is being eaten up rapidly according to df however du numbers really never change. lsof and fuser cannot see any unlinked files on either the NFS server or remote clients. ... (1 Reply)
Discussion started by: masterpengu
1 Replies
NFSSERVER(8)						      System Manager's Manual						      NFSSERVER(8)

NAME
nfsserver, portmapper, pcnfsd, 9auth - NFS service SYNOPSIS
aux/nfsserver [ rpc-options... ] [ nfs-options... ] aux/pcnfsd [ rpc-options... ] aux/portmapper [ rpc-options... ] 9auth [ auth-options ] user DESCRIPTION
These programs collectively provide NFS access to Plan 9 file servers. Nfsserver, pcnfsd, and portmapper run on a Plan 9 CPU server, and should be started in that order. 9auth is run by a user on a client machine--usually a Unix system, never Plan 9--for optional per-user authentication. By default, all users on client machines have the access privileges of the Plan 9 user The rpc-options are all intended for debugging: -r Reject: answer all RPC requests by returning the AUTH_TOOWEAK error. -v Verbose: show all RPC calls and internal program state, including 9P messages. (In any case, the program creates a file /srv/name.chat where name is that of the program; echoing or into this file sets or clears the -v flag dynamically.) -D Debug: show all RPC messages (at a lower level than -v). This flag may be repeated to get more detail. The nfs-options are: -a addr Set up NFS service for the 9P server at network address addr. -f file Set up NFS service for the 9P server at file (typically an entry in /srv). -n Do not allow per-user authentication. -c file File contains the uid/gid map configuration. It is read at startup and subsequently every hour (or if is echoed into /srv/nfsserver.chat). Blank lines or lines beginning with are ignored; lines beginning with are executed as commands; otherwise lines contain four fields separated by white space: a regular expression (in the notation of regexp(6)) for a class of servers, a regular expression for a class of clients, a file of user id's (in the format of a Unix password file), and a file of group id's (same format). NFS clients must be in the Plan 9 /lib/ndb database. The machine name and IP address given in the NFS mount request must match the entry in the database. Pcnfsd is a toy program that authorizes PC-NFS clients. All clients are mapped to uid=1, gid=1 (daemon on most systems) regardless of name or password. The options for 9auth are: -D Debug flag. -r root Authenticate to the file system at root (default /n/bootes). -d Delete the authorization for user. In the absence of -d, 9auth will present a challenge to be encrypted with the user's Plan 9 password (using netkey, for example; see passwd(1)). If the response is correct, subsequent NFS transactions will take place with the user's privileges. EXAMPLES
A simple /lib/ndb/nfs might contain: !9fs tcp!ivy .+ [^.]+.cvrd.hall.edu /n/ivy/etc/passwd /n/ivy/etc/group A typical entry in /rc/bin/cpurc might be: aux/nfsserver -a il!bootes -a il!fornax -c /lib/ndb/nfs aux/pcnfsd aux/portmapper Assuming the cpu server's name is eduardo, the mount commands on the client would be: /etc/mount -o soft,intr eduardo:bootes /n/bootes /etc/mount -o soft,intr eduardo:fornax /n/fornax Note that a single instance of nfsserver may provide access to several 9P servers. FILES
/lib/ndb/nfs List of uid/gid maps. /sys/log/nfs Log file. SOURCE
/sys/src/cmd/service/nfs /sys/src/cmd/unix/9auth.c SEE ALSO
Netkey in passwd(1), regexp(6), u9fs(4) RFC1057, RPC: Remote Procedure Call Protocol Specification, Version 2. RFC1094, NFS: Network File System Protocol Specification. NFSSERVER(8)
All times are GMT -4. The time now is 10:02 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy