[Gluster-users] root_squash problem

[Laurent Bardi]

Hi,

i ve setup a gluster mirror (2 machine runing debian jessie, and gluster
rebuild from source 3.8.3-1).

the undelying brick is on xfs filesystem (i ve mounted it with noattr2)

i ve directories in gluster vols with posix acl : fine

i share them with the native gluster NFS : fine if i set no_root_squash.

if i set root squash : the nfs.log says :

[2016-09-12 06:50:30.841786] W [MSGID: 114031]
[client-rpc-fops.c:2930:client3_3_lookup_cbk]
0-vnas_gluster_users-client-1: remote operation failed. Path:
<gfid:a52bb061-9e2d-4603-8471-454b972fff04>/tetutetu1234
(00000000-0000-0000-0000-000000000000) [Permission non accordée]
[2016-09-12 06:50:30.842721] W [MSGID: 114031]
[client-rpc-fops.c:2930:client3_3_lookup_cbk]
0-vnas_gluster_users-client-0: remote operation failed. Path:
<gfid:a52bb061-9e2d-4603-8471-454b972fff04>/tetutetu1234
(00000000-0000-0000-0000-000000000000) [Permission non accordée]
[2016-09-12 06:50:30.842793] E [MSGID: 112198]
[nfs3-helpers.c:3647:nfs3_fh_resolve_entry_lookup_cbk] 0-nfs-nfsv3:
Lookup failed: <gfid:a52bb061-9e2d-4603-8471-454b972fff04>/tetutetu1234:
Permission non accordée [Permission non accordée]
[2016-09-12 06:50:30.842825] E [MSGID: 112069]
[nfs3.c:2838:nfs3_mkdir_resume] 0-nfs-nfsv3: Permission non accordée:
(10.1.1.16:1003) vnas_gluster_users : a52bb061-9e2d-4603-8471-454b972fff04

I ve tries many things (such as invalidate glusternfs, replace it with
kernel-nfs and it works). But  i need to use the gluster nfs as i want
to use CTDB (from samba) too , in order to build a real active/active
data cluster.

Moreover i need to use root_squashing too because it will be the
fileserver for all machine in the institute, so if someone with a linux
laptop and root access can mount the nfs gluster  homedirs it will have
acces to all homedirs : no security...

I ve not took the option nfs-ganesha as : i need to have NFS v3 with
posix acl (old machines such as irix)


A) Is there a way to increase the gluster nfs log verbosity ?

b) i ve in the nfs.log a lot of :
[2016-09-12 06:56:59.775860] E [MSGID: 112195]
[nfs-fops.c:84:nfs_fix_groups] 0-nfs-server: getpwuid_r(118) found nothing
The message "E [MSGID: 112195] [nfs-fops.c:84:nfs_fix_groups]
0-nfs-server: getpwuid_r(118) found nothing" repeated 74 times between
[2016-09-12 06:56:59.775860] and [2016-09-12 06:57:16.373337]

uid=118 is snmp ?Should i worry about it ?

C) Is there another solution with this damned root_squash ?

D) Am I totally insane to try this :): ?

Many thanks in advance


-- 
Laurent BARDI /  RSI CNRS-IPBS / CRSSI DR14
INSTITUT  de PHARMACOLOGIE et de BIOLOGIE STRUCTURALE
Tel : 05-61-17-59-05    http://www.ipbs.fr/
Fax : 05-61-17-59-94  	Laurent.BardiATipbs.fr
CNRS-IPBS 205 Route de Narbonne 31400 TOULOUSE FRANCE
...
J'étais indéniablement misanthrope.
Je voulus traverser à gué un marigot infesté d'imbéciles. 
Quand j'atteignis l'autre rive, j'étais devenu philanthrope.