Hi all, I installed Lustre 1.4.6 with patched Linux kernel on RHE2.6.9 and I ran the test script NAME=local sh llmount.sh which is located in the tests directory, it executed successfully and give the mount path with /mnt/lustre. I can able to remove the mount using ''NAME=local sh llmountcleanup.sh '' and remount successfully. But when I try to setup Client, MDS, and two OSTs on a single node, I got the following error loading module: libcfs srcdir None devdir libcfs Bad module options? Check dmesg. ! modprobe (error 1):> FATAL: Module libcfs not found.How can I solve this?? The script is as follows #!/bin/sh # local.sh # Create node rm -f local.xml lmc -m local.xml --add node --node dev1 lmc -m local.xml --add net --node dev1 --nid dev1 --nettype tcp # Configure MDS lmc -m local.xml --format --add mds --node dev1 --mds mds-test --fstype ext3 --dev /tmp/mds-test --size 50000 # Configure OSTs lmc -m local.xml --add lov --lov lov-test --mds mds-test --stripe_sz 1048576 --stripe_cnt 0 --stripe_pattern 0 lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost1-test --fstype ext3 --dev /tmp/ost1-test --size 100000 lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost2-test --fstype ext3 --dev /tmp/ost2-test --size 100000 # Configure client lmc -m local.xml --add mtpt --node dev1 --path /mnt/lustre --mds mds-test --lov lov-test Thanks, Prajith -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060621/5deff0f8/attachment.html
On Jun 21, 2006 19:34 +0530, Prajith Lal wrote:> I installed Lustre 1.4.6 with patched Linux kernel on RHE2.6.9 and I > ran the test script NAME=local sh llmount.sh which is located in the > tests directory, it executed successfully and give the mount path with > /mnt/lustre. I can able to remove the mount using ''NAME=local sh > llmountcleanup.sh '' and remount successfully. But when I try to setup > Client, MDS, and two OSTs on a single node, I got the following error > loading module: libcfs srcdir None devdir libcfs > Bad module options? Check dmesg.Did you check dmesg?> ! modprobe (error 1): > >FATAL: Module libcfs not found. > How can I solve this?? > The script is as follows > #!/bin/sh > > # local.sh > > # Create node > rm -f local.xml > lmc -m local.xml --add node --node dev1 > lmc -m local.xml --add net --node dev1 --nid dev1 --nettype tcp > > # Configure MDS > lmc -m local.xml --format --add mds --node dev1 --mds mds-test > --fstype ext3 --dev /tmp/mds-test --size 50000 > > # Configure OSTs > lmc -m local.xml --add lov --lov lov-test --mds mds-test --stripe_sz > 1048576 --stripe_cnt 0 --stripe_pattern 0 > lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost1-test > --fstype ext3 --dev /tmp/ost1-test --size 100000 > lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost2-test > --fstype ext3 --dev /tmp/ost2-test --size 100000 > > # Configure client > lmc -m local.xml --add mtpt --node dev1 --path /mnt/lustre --mds > mds-test --lov lov-testYour script seems fine otherwise. The other option is to use the "lov.sh" script as a starting point, which by default configures 5 OSTs, though this is tunable. As above, run "NAME=lov sh llmount.sh". Cheers, Andreas -- Andreas Dilger Principal Software Engineer Cluster File Systems, Inc.
Hi, Thanks for your reply. I resolve the problem. The thing is after patching the kernel modules I forget to run the ''depmod'' command. After executing the command, it works fine . Now I am trying to setup Lustre with multiple nodes. I''ve some doubt 1. Is the Lustre filesystem supports data locality? 2. How will I access a file using Java which is stored in Lustre FS? Thanks, Prajith On 6/22/06, Andreas Dilger <adilger@clusterfs.com> wrote:> > On Jun 21, 2006 19:34 +0530, Prajith Lal wrote: > > I installed Lustre 1.4.6 with patched Linux kernel on RHE2.6.9 and I > > ran the test script NAME=local sh llmount.sh which is located in the > > tests directory, it executed successfully and give the mount path with > > /mnt/lustre. I can able to remove the mount using ''NAME=local sh > > llmountcleanup.sh '' and remount successfully. But when I try to setup > > Client, MDS, and two OSTs on a single node, I got the following error > > loading module: libcfs srcdir None devdir libcfs > > Bad module options? Check dmesg. > > Did you check dmesg? > > > ! modprobe (error 1): > > >FATAL: Module libcfs not found. > > How can I solve this?? > > The script is as follows > > #!/bin/sh > > > > # local.sh > > > > # Create node > > rm -f local.xml > > lmc -m local.xml --add node --node dev1 > > lmc -m local.xml --add net --node dev1 --nid dev1 --nettype tcp > > > > # Configure MDS > > lmc -m local.xml --format --add mds --node dev1 --mds mds-test > > --fstype ext3 --dev /tmp/mds-test --size 50000 > > > > # Configure OSTs > > lmc -m local.xml --add lov --lov lov-test --mds mds-test --stripe_sz > > 1048576 --stripe_cnt 0 --stripe_pattern 0 > > lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost1-test > > --fstype ext3 --dev /tmp/ost1-test --size 100000 > > lmc -m local.xml --add ost --node dev1 --lov lov-test --ost ost2-test > > --fstype ext3 --dev /tmp/ost2-test --size 100000 > > > > # Configure client > > lmc -m local.xml --add mtpt --node dev1 --path /mnt/lustre --mds > > mds-test --lov lov-test > > Your script seems fine otherwise. The other option is to use the > "lov.sh" script as a starting point, which by default configures 5 > OSTs, though this is tunable. As above, run "NAME=lov sh llmount.sh". > > Cheers, Andreas > -- > Andreas Dilger > Principal Software Engineer > Cluster File Systems, Inc. > >-- Lals -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060622/0ed59245/attachment.html
On Jun 23, 2006 10:18 +0530, Prajith Lal wrote:> I resolve the problem. The thing is after patching the kernel modules I > forget to run the ''depmod'' command. After executing the command, it works > fine . Now I am trying to setup Lustre with multiple nodes. I''ve some doubt > > 1. Is the Lustre filesystem supports data locality?I don''t understand what this means.> 2. How will I access a file using Java which is stored in Lustre FS?The same way you would access a file from Java on any other filesystem. Cheers, Andreas -- Andreas Dilger Principal Software Engineer Cluster File Systems, Inc.
Hi, Thanks for spending your time for me. Data locality means if we are storing a large amount of data in lustre filesystem, can we identify it directly? means can we access it directly as in nfs. Thanks, Prajith On 6/23/06, Andreas Dilger <adilger@clusterfs.com > wrote:> > On Jun 23, 2006 10:18 +0530, Prajith Lal wrote: > > I resolve the problem. The thing is after patching the kernel modules I > > forget to run the ''depmod'' command. After executing the command, it > works > > fine . Now I am trying to setup Lustre with multiple nodes. I''ve some > doubt > > > > 1. Is the Lustre filesystem supports data locality? > > I don''t understand what this means. > > > 2. How will I access a file using Java which is stored in Lustre FS? > > The same way you would access a file from Java on any other filesystem. > > > Cheers, Andreas > -- > Andreas Dilger > Principal Software Engineer > Cluster File Systems, Inc. > >-- Lals -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20060623/d987b05a/attachment.html