Unsolved
This post is more than 5 years old
10 Posts
0
1575
One virtual disk with two iscsi initiators connection on AX 4-5i
Dear,
I have a AX4-5i and configure only one virtual disk to assigned two iscsi initiators/server. My plan was to access one disk using two iscsi server expecting to see the same file system. I managed to mount a disk on two server as /dellarray. However, when I write on disk I realize I am not see the same file system. It appeared to show athat they do not share the file system although PowerPath showed the same device ID. Where I should be focusing to resolve this? My first question is though that whether this is even possible. I am using Navisphere extress for EMC storaage side and I am using two iscsi initiator from linux redhat 5. Thanks.
Regards,
Taeksu Shin
MorbidAngel1
62 Posts
1
August 22nd, 2011 18:00
Hi,
Did you try mounting the File system again after writing on node 1.It should see the same data on node 2 after remounting.
tshin1
10 Posts
0
August 23rd, 2011 06:00
Dear Arifur,
Here is what did.
node1 and node2 is assigned to the same virtual disk, vd1.
At node1,
node1> mkfs.xfs –f /dev/emcpowerc
meta-data=/dev/emcpowerc isize=256 agcount=32, agsize=16777216 blks
= sectsz=512 attr=0
data = bsize=4096 blocks=536870912, imaxpct=25
= sunit=0 swidth=0 blks, unwritten=1
naming =version 2 bsize=4096
log =internal log bsize=4096 blocks=32768, version=1
= sectsz=512 sunit=0 blks, lazy-count=0
realtime =none extsz=4096 blocks=0, rtextents=0
node2> mkfs.xfs /dev/emcpowera
mkfs.xfs: /dev/emcpowera appears to contain an existing filesystem (xfs).
mkfs.xfs: Use the -f option to force overwrite.
node2> mkfs.xfs –f /dev/emcpowera
meta-data=/dev/emcpowera isize=256 agcount=32, agsize=16777216 blks
= sectsz=512 attr=0
data = bsize=4096 blocks=536870912, imaxpct=25
= sunit=0 swidth=0 blks, unwritten=1
naming =version 2 bsize=4096
log =internal log bsize=4096 blocks=32768, version=1
= sectsz=512 sunit=0 blks, lazy-count=0
realtime =none extsz=4096 blocks=0, rtextents=0
node1> mount –xfs /dev/emcpowerc /dellarray
node2> cd /dellarray
node1> dd if=/dev/zero of=node1.img bs=8k count=256k
node1> ls
node1.img
node2> mount –xfs /dev/emcpowera /dellarray
node2> cd /dellarray
node2> dd if=/dev/zero of=node2.img bs=8k count=256k
node2> ls
node2.img
Here I do not see the node1.img which generated at node1.
I am suspicious on the mkfs.xfs –f options but without –f option, mkfs does not works.
Thanks.
Regards,
Taeksu
tshin1
10 Posts
0
August 23rd, 2011 11:00
Dear Arifur,
Your statement is correct. After remounting of disk at node1, I can see the file created at the node2.
So this indicates one virtual disk I/O shared by two server needs to have some kind of disk clustering
software to see the same file structure?
Regards,
Taeksu
MorbidAngel1
62 Posts
1
August 23rd, 2011 17:00
Yes...You are right.It should be maintained by some File System Service(NFS,GFS) or Cluster Service.....
kelleg
4.5K Posts
1
August 29th, 2011 06:00
Whenever you attache more than one server to the same disk (or LUN), you MUST have some type of clustering software to arbitrate who has access to the data at any one time otherwise you will corrupt the data on the disk.
Please remember to mark the question as Answered when you get the correct or best answer. Also, please award points to the person providing the best answer.
glen
tshin1
10 Posts
0
August 29th, 2011 12:00
Thanks for updating on this.
Regards,
Taeksu
SKT2
1.3K Posts
0
September 17th, 2011 18:00
what version/update of rhel5 are you using; after a certain version level you get teh red-hat cluster suite along with the OS ( like bundled and free with os) . In this cas you may share them and see the data, but needs to be used by only one node at a time. For concurrent access you will needs GFS,OCFS like features