The Lustre manual says, "Use the proper block device (sgX versus sdX) ...
sgX is a special interface that bypasses the block layer and the I/O
scheduler..." Does this mean that we are supposed to use it after the --dev
option?
Normally, my lmc script includes a line like:
lmc -m demo192.xml --add ost --node roger-tyan --lov lov-ts --ost ost1-ts
--fstype ext3 --dev /dev/sdb
I changed that to say --dev /dev/sg1 .
I tried this, then I ran the command:
lconf -v --abort_recovery --reformat --node roger-tyan demo192.xml
This gave the error:
size of loopback file '/dev/sg1' must be larger than 8MB, but is set to 0
So, I modified the lmc command to read:
lmc -m demo192.xml --add ost --node roger-tyan --lov lov-ts --ost ost1-ts
--fstype ext3 --dev /dev/sg1 --size 2048000
Then, lconf just failed. What am I doing wrong?
By the way, I'm running Lustre 1.4.6.4, and my whole lmc script is now:
lmc -m demo192.xml --add net --node roger-ha-3 --nid [EMAIL PROTECTED]
--nettype tcp
lmc -m demo192.xml --add net --node roger-tyan --nid [EMAIL PROTECTED]
--nettype tcp
lmc -m demo192.xml --add net --node client --nid * --nettype tcp
lmc -m demo192.xml --add mds --node roger-ha-3 --mds mds192 --fstype ext3
--dev /tmp/MDS --size 50000
lmc -m demo192.xml --add lov --lov lov-ts --mds mds192 --stripe_sz 1048576
--stripe_cnt 0 --stripe_pattern 0
lmc -m demo192.xml --add ost --node roger-tyan --lov lov-ts --ost ost1-ts
--fstype ext3 --dev /dev/sg1 --size 2048000
lmc -m demo192.xml --add mtpt --node client --path /mnt/lustre192 --mds
mds192 --lov lov-ts
-Roger
_________________________________________________________________
Get in the mood for Valentine's Day. View photos, recipes and more on your
Live.com page.
http://www.live.com/?addTemplate=ValentinesDay&ocid=T001MSN30A0701
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
