Discussion:
[Gluster-users] Gluster client
Alfredo De Luca
2018-10-15 19:33:14 UTC
Permalink
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0

But what happened when NODE1 is unavailable?

Clients are centos 7.5 so the servers

Thanks
--
*Alfredo*
Diego Remolina
2018-10-15 19:39:55 UTC
Permalink
You may have a typo, "_netdev" you are missing "_".

Give that a try.

Diego
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
*Alfredo*
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
Alfredo De Luca
2018-10-15 19:41:13 UTC
Permalink
Hi Diego...sorry it's a typo here on the email... but I ve put _netdev in
the fstab.

Thanks
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
*Alfredo*
--
*Alfredo*
Dmitry Melekhov
2018-10-16 04:28:09 UTC
Permalink
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them
they don't  mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
Dunno, we run gluster on the same nodes as VM, so we put localhost in
domain definitions.
In your situation I'd use something like VRRP ( keepalived , for instance).
Post by Alfredo De Luca
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
/*Alfredo*/
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
Alfredo De Luca
2018-10-16 12:38:37 UTC
Permalink
ok.. found out that backupvolfile-server is correct on fstab in case at
mounting time the primary server is not responding..... so
the backupvolfile-server will fail to the next server.
Also it seems that if you put <server1:/vol0> in your fstab and the server1
fails during normal operation the client will fail to the next one in the
cluster.

Cheers
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
Dunno, we run gluster on the same nodes as VM, so we put localhost in
domain definitions.
In your situation I'd use something like VRRP ( keepalived , for instance).
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
*Alfredo*
_______________________________________________
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
Dmitry Melekhov
2018-10-16 12:40:41 UTC
Permalink
ok.. found out that backupvolfile-server is correct on fstab in case
at mounting time the primary server is not responding..... so
the backupvolfile-server will fail to the next server.
Also it seems that if you put <server1:/vol0> in your fstab and the
server1 fails during normal operation the client will fail to the next
one in the cluster.
Cheers
I thought you asking about initial mount if server one is not available :-)
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am
a bit newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot
them they don't mount it automatically
2. Not sure what to exactly put in the fastab as right now
someone had <node1>:/vol1 /volume1 glusterfs default,netdev 0 0
Dunno, we run gluster on the same nodes as VM, so we put localhost
in domain definitions.
In your situation I'd use something like VRRP ( keepalived , for instance).
Post by Alfredo De Luca
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
/*Alfredo*/
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
/*Alfredo*/
Alfredo De Luca
2018-10-16 12:44:08 UTC
Permalink
thanks Dmitry...maybe I wasn't clear enough....but I was asking for both.
Anyway it should be fine now.

Cheers
Post by Alfredo De Luca
ok.. found out that backupvolfile-server is correct on fstab in case at
mounting time the primary server is not responding..... so
the backupvolfile-server will fail to the next server.
Also it seems that if you put <server1:/vol0> in your fstab and the
server1 fails during normal operation the client will fail to the next one
in the cluster.
Cheers
I thought you asking about initial mount if server one is not available :-)
Post by Alfredo De Luca
Hi all.
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
Dunno, we run gluster on the same nodes as VM, so we put localhost in
domain definitions.
In your situation I'd use something like VRRP ( keepalived , for instance).
But what happened when NODE1 is unavailable?
Clients are centos 7.5 so the servers
Thanks
--
*Alfredo*
_______________________________________________
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
--
*Alfredo*
Stefan Kania
2018-10-16 12:45:49 UTC
Permalink
I wrote my own systemd-script:
------------------
[Unit]
Description = Data dir
After=network.target glusterfs-server.service
Required=network-online.target

[Mount]
What=knoten-1:/gv1
Where=/glusterfs
Type=glusterfs
Options=defaults,acl

[Install]
WantedBy=multi-user.target
------------------
This script must be in /etc/systemd/system and the name must be
<mountpoint>.mount
Hi all. 
I have 3 nodes glusterfs servers and multiple client and as I am a bit
newbie on this not sure how to setup correctly the clients.
1. The clients mounts the glusterfs in fstab but when I reboot them they
don't  mount it automatically
2. Not sure what to exactly put in the fastab as right now someone had
<node1>:/vol1 /volume1 glusterfs default,netdev 0 0
But what happened when NODE1 is unavailable?
The Client will only get a list of all hosts in the cluster in if one
node is down the client will take another node.
Clients are centos 7.5 so the servers
Thanks
--
/*Alfredo*/
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
Dave Sherohman
2018-10-16 13:19:01 UTC
Permalink
Post by Stefan Kania
Post by Alfredo De Luca
But what happened when NODE1 is unavailable?
The Client will only get a list of all hosts in the cluster in if one
node is down the client will take another node.
You have to connect to the volume first before you can get that list
from the server. What happens if, when you make the initial connection,
you try to connect to a node that's down? I would certainly expect it
to fail, since the client doesn't have that list yet, so it has no idea
what other nodes it might attempt to connect to.

I primarily use gluster for VM disk images, so, in my case, I list all
the gluster nodes in the VM definition and, if the first one isn't
reachable, then it tries the second and so on until it finds one that's
available to connect to.

What's the fstab equivalent?
--
Dave Sherohman
j***@mail.de
2018-10-16 13:29:25 UTC
Permalink
Post by Dave Sherohman
What's the fstab equivalent?
Hi,

with the fuse client you can try the mount option backup-volfile-servers=server1:server2
This gives alternate points to ask for the volume info.

Joachim
-------------------------------------------------------------------------------------------------
FreeMail powered by mail.de - MEHR SICHERHEIT, SERIOSITÄT UND KOMFORT
Alfredo De Luca
2018-10-16 14:00:39 UTC
Permalink
The client was already connected to the volume so it takes the info about
the nodes. I think I need to add anyway the backup-volfile-servers in the
fstab so it can check at boot time.

Cheers
Post by j***@mail.de
Post by Dave Sherohman
What's the fstab equivalent?
Hi,
with the fuse client you can try the mount option
backup-volfile-servers=server1:server2
This gives alternate points to ask for the volume info.
Joachim
-------------------------------------------------------------------------------------------------
FreeMail powered by mail.de - MEHR SICHERHEIT, SERIOSITÄT UND KOMFORT
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
Vlad Kopylov
2018-10-16 22:05:05 UTC
Permalink
You can add fetch-attempts=5 to fstab, so it will try to connect more,
never had an issue after this

Problem might be as it might connect to the other server not the local one,
starting to push all reads through the network - so close client ports on
other nodes but to local

v
Post by Alfredo De Luca
The client was already connected to the volume so it takes the info about
the nodes. I think I need to add anyway the backup-volfile-servers in the
fstab so it can check at boot time.
Cheers
Post by j***@mail.de
Post by Dave Sherohman
What's the fstab equivalent?
Hi,
with the fuse client you can try the mount option
backup-volfile-servers=server1:server2
This gives alternate points to ask for the volume info.
Joachim
-------------------------------------------------------------------------------------------------
FreeMail powered by mail.de - MEHR SICHERHEIT, SERIOSITÄT UND KOMFORT
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
Alfredo De Luca
2018-10-17 15:29:48 UTC
Permalink
What does fetch-attempts=5 do?
Post by Vlad Kopylov
You can add fetch-attempts=5 to fstab, so it will try to connect more,
never had an issue after this
Problem might be as it might connect to the other server not the local
one, starting to push all reads through the network - so close client ports
on other nodes but to local
v
Post by Alfredo De Luca
The client was already connected to the volume so it takes the info about
the nodes. I think I need to add anyway the backup-volfile-servers in
the fstab so it can check at boot time.
Cheers
Post by j***@mail.de
Post by Dave Sherohman
What's the fstab equivalent?
Hi,
with the fuse client you can try the mount option
backup-volfile-servers=server1:server2
This gives alternate points to ask for the volume info.
Joachim
-------------------------------------------------------------------------------------------------
FreeMail powered by mail.de - MEHR SICHERHEIT, SERIOSITÄT UND KOMFORT
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
Vlad Kopylov
2018-10-18 21:20:37 UTC
Permalink
Maximum number of connect attempts to server
Post by Alfredo De Luca
What does fetch-attempts=5 do?
Post by Vlad Kopylov
You can add fetch-attempts=5 to fstab, so it will try to connect more,
never had an issue after this
Problem might be as it might connect to the other server not the local
one, starting to push all reads through the network - so close client ports
on other nodes but to local
v
On Tue, Oct 16, 2018 at 10:01 AM Alfredo De Luca <
Post by Alfredo De Luca
The client was already connected to the volume so it takes the info
about the nodes. I think I need to add anyway the backup-volfile-servers
in the fstab so it can check at boot time.
Cheers
Post by j***@mail.de
Post by Dave Sherohman
What's the fstab equivalent?
Hi,
with the fuse client you can try the mount option
backup-volfile-servers=server1:server2
This gives alternate points to ask for the volume info.
Joachim
-------------------------------------------------------------------------------------------------
FreeMail powered by mail.de - MEHR SICHERHEIT, SERIOSITÄT UND KOMFORT
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
_______________________________________________
Gluster-users mailing list
https://lists.gluster.org/mailman/listinfo/gluster-users
--
*Alfredo*
Loading...