Hi,
Thanks for you answers. Please let me ask another questions.
1) Are there any recommendations regarding size of block device for
mgs ?
2) Also do I need to have any raid configuration for mgs block device.
3) If I would lost mgs device what are the steps to recreate it and
get access to filesystem
Thanks You
Wojciech
On 17 Aug 2007, at 17:37, Anand Bisen wrote:
Wojciech,
For question 2: it's a typo. it should be /dev/sda1 and not /dev/
sda (or if you are using the whole device then vice versa, both
would work).
For question 1: I think either the MGS should be on a separate
machine and if MDT and MGS are co-located on a single node, i don't
think it would make a big difference if they are on different
partitions (sda1, sda2) or not. Although depending upon the total
number of client's and their frequency of mount's/umount's it might
help keeping those IO's on a different block device (sda, sdb).
Thanks
Anand
On Aug 17, 2007, at 8:11 AM, Wojciech Turek wrote:
On 17 Aug 2007, at 16:04, Wojciech Turek wrote:
Hi,
I have two short questions:
1) Is it better to configure MGS on different partition then MDT?
For example:
I have node01 which has two partitions /dev/sda1 and /dev/sda2
I am planning to create MGS on /dev/sda1 and MDT on /dev/sda2
2) If above is actually possible? I read in manual follwing:
3 Start Lustre on the MGS node.
$ mkfs.lustre --mgs /dev/sda
4 Make a mount point on MGS for the file system and mount it.
$ mkdir -p /mnt/mgs
$ mount -t lustre /dev/sda1 /mnt/mgs
It looks like I will have to create mgs on entire device /dev/sda
and then mount it using /dev/sda1 Is that right or typo?
Regards
Wojciech Turek
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
Mr Wojciech Turek
Assistant System Manager
University of Cambridge
High Performance Computing service
email: [EMAIL PROTECTED]
tel. +441223763517
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
