Details
-
Improvement
-
Resolution: Duplicate
-
Minor
-
None
-
Lustre 2.10.5
-
None
-
Server:
{noformat}
[root@oss01 lustre-2.10.5_ddn7]# uname -a
Linux le-oss01 3.10.0-862.9.1.el7_lustre.ddn1.x86_64 #1 SMP Tue Sep 11 19:05:37 JST 2018 x86_64 x86_64 x86_64 GNU/Linux
{noformat}
{noformat}
lustre-2.10.5_ddn7-1.el7.x86_64
lustre-osd-ldiskfs-mount-2.10.5_ddn7-1.el7.x86_64
{noformat}
Client:
{noformat}
[root@a2-029 ~]# rpm -qa | grep ^lustre
lustre-client-2.10.5_ddn2_2_gda48415-1.el7.x86_64
[root@lo-a2-029 ~]# uname -a
Linux lo-a2-029 3.10.0-862.14.4.el7.x86_64 #1 SMP Wed Sep 26 15:12:11 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
{noformat}Server: {noformat} [ root@oss01 lustre-2.10.5_ddn7]# uname -a Linux le-oss01 3.10.0-862.9.1.el7_lustre.ddn1.x86_64 #1 SMP Tue Sep 11 19:05:37 JST 2018 x86_64 x86_64 x86_64 GNU/Linux {noformat} {noformat} lustre-2.10.5_ddn7-1.el7.x86_64 lustre-osd-ldiskfs-mount-2.10.5_ddn7-1.el7.x86_64 {noformat} Client: {noformat} [ root@a2-029 ~]# rpm -qa | grep ^lustre lustre-client-2.10.5_ddn2_2_gda48415-1.el7.x86_64 [ root@lo-a2-029 ~]# uname -a Linux lo-a2-029 3.10.0-862.14.4.el7.x86_64 #1 SMP Wed Sep 26 15:12:11 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux {noformat}
-
9223372036854775807
Description
We're currently doing some tests in a multi-tenancy environment with Lustre routers. In this environment we have defined several filesets exported to the same clients but with different LNETs, like:
Client LNETs: tcp0, tcp1
Server LNETs: tcp10, tcp11
The Lustre routers connect the LNETs in the following way: tcp0 <> tcp10, tcp1 <> tcp11
- Fileset data1: ranges=<client_ip>@tcp0
- Fileset data2: ranges=<client_ip>@tcp1
Now when we look at the 'network' option in mount.lustre:
network=net Limit connections from the client to be on the network NID specified by 'net'. ´net´ designates a single network NID, like 'o2ib2' or 'tcp1'. This option can be useful in case of several Lustre client mount points on the same node, with each mount point using a different network. It is also interesting when running Lustre clients from containers, by restricting each container to a specific network.
Since it's limiting the connections from the client we should use the following command to mount dataset2:
mount -t lustre -o network=tcp1 <mgs_ip>@tcp11:/ /lustre/dataset2
But that's not working because network needs to be the network on the server. This is working:
mount -t lustre -o network=tcp11 <mgs_ip>@tcp11:/ /lustre/dataset2
The semantics of the option is not respected.
After discussing with Sebastien it looks like LU-11959 would be the patch for this.
Attachments
Issue Links
- duplicates
-
LU-11959 Improve handling of 'network' client mount option
-
- Open
-
ok thanks!