Newbie to Celerra/NAS and have questions

Welcome to the Celerra World

1. The RAID Groups / LUNs / Storage groups are just the same as connecting any SAN host to a CLARiiON - so from CLARiiON stand point these needs to be done as usual with some special consideration for Celerra. The RAID groups need to be of supported type with specific disks - say for example RAID 5 with 5 (4+1) or 9 (8+1) disks, RAID 6 with 4+2 or 6+2 or 12+2 , RAID 1 etc - it also depends on whether you are creating it on FC disks or SATA disks. The number of LUNs are typically either 2 or 4 in a RAID Group.

If this Celerra is an Integrated (or Non-FC) one that means the backend CLARiiON is all for the NAS directly connected, then Celerra configures the Boot LUNs during the installation and the rest of the data LUNs are configured through setup_clariion script or using Celerra Manager GUI.

If this a gateway NAS or FC-enabled one, you need to configure CLARiiON LUNs accordingly - add them to the Celerra Storage Group. All the data movers on the NAS needs to be manually registered and added to the storage group.

Please refer to the following links which may provide more details -

http://corpusweb130.corp.emc.com/ns/common/post_install/Config_storage.htm

http://corpusweb130.corp.emc.com/ns/common/post_install/Config_storage_integrated.htm

http://corpusweb130.corp.emc.com/ns/common/post_install/Config_storage_FC_NaviManager.htm

Now comes the Pools, Volumes & File systems on Celerra. Each LUN on the backend CLARiiON when presented to the Celerra - are treated as a disk device (dxx) on the Celerra. If you are using supported RAID Group and LUN configuration, all identical LUNs add the space to a system defined pool - like all 4+1 RAID 5 LUNs space is under clar_r5_performance pool or all 8+1 RAID5 LUNs are added as clar_r5_economy pool as so on.

When you want to create File system, most of the cases, it is recommended to use AVM (Automated Volume management) on the Celerra, so that you need not to worry about any more volume management in the form of creating slice, stripe, metas etc. AVM does it with a very vast and powerful algorithm depending on the configuration and file system size. You need to simply choose the name & size of the file system you want to create and choose the storage pool from where you want to create. By default slicing is enabled, so keep it like that.

Advanced users may opt for Manual Volume management if absolutely needed for the requirement.

I 'll suggest to take a look at the following documents on powerlink under the Navigation path -

Home > Support > Technical Documentation and Advisories > Hardware/Platforms Documentation > Celerra Network Server > Maintenance/Administration


"Implementing Automatic Volume Management with Celerra"

"Managing Celerra Volumes and File Systems with Automatic Volume Management"

"Managing Celerra Volumes and File Systems Manually "

You may also find some more useful documents on the Powerlink -
Home > Support > Technical Documentation and Advisories > Hardware/Platforms Documentation > Celerra Network Server


2. It is highly not recommended to span a file system across multiple pools. In many cases this is not possible, if the pools are on different disk types or RAIDs - but whenever possible (say 4+1 RAID 5 and 8+1 RAID 5) - it is still not recommended to mix them from performance point of view unless absolutely needed. When a pool is almost at capacity, the recommendation will be to add more disks and create new RG/LUNs to extend the pool capacity - or if another pool is available with more capacity, create a larger file system there and migrate from the existing one.

I realize it has been such a long post - sorry for that, hope this helps and I am sure others will have more valuable inputs.

Thanks,
Sandip

You may create a new file system of desired size on the economy pool - create the new share or export and present the same to end user where you can perform a host based migration say using robocopy or emcopy etc.

Another option may be to use Celerra Replicator with in the box to create a new file system on the economy pool and the replicate the existing file system to it - once the cutover is done, you may extend the new file system. Depending on the NAS code version you are using, it will use fs_copy or nas_copy. If you can take the file system offline - only one fs_copy or nas_copy can migrate the data or else, you need to take care of the incremental changes and follow replication process.

However, since you mentioned the economy pool is not used at all - there may be possibility to reconfigure the CLARiiON RAID Groups from 8+1 RAID 5 to 4+1 RAID 5 - at least for few Raid groups. Can you please provide the details of the CLARiiON RAID Group and LUN configuration used by Celerra - it may be bit helpful to suggest.

Thanks,
Sandip

Okay, I think the light is beginning to get brighter.

The economy pool is set up with two Raid 5 (8+1) groups of FC disk. The performance pool is Raid 5 (4+1). The two 8+1 groups consist of two LUN's each.

I hope that is the answer that you are looking for, if not, I will make another attempt.

Thanks so much again.

J

Wow, that is great information. Thanks.

A followup question if I can. The situation we have gotten ourselves into is this:

The CLARiiON back end is at capacity as far as disk is concerned.

The clar_r5_performance pool is at 95% capacity. Our clar_r5_economy pool is empty.

What would be your recommendation on migrating/moving/adding space from the empty pool to be used by file systems in the clar_r5_performance pool?

Thanks again.

J

Thanks - it will be even clearer to know the bus_enclo_disk numbers for all those RAID groups. And also any HS if any, in the same enclosure.

I understand you have other hosts connected to the same CLARiiON - for which you have different RAID groups.

What I was suggesting as may be one possibility is to destroy the 8+1 RAID groups and instead build 4+1 RAID groups, then follow the procedure to add new LUNs on the newly created 4+1 RAID 5 groups.

It depends on the drive location thus - and you can not create a 4+1 RAID group for Celerra spanning across enclosure.

Also, keep in mind that you need to identify the disks on the Celerra which are part of the economy pool and delete these disks on Celerra (use nas_disks - delete -perm command) before taking out those LUNs from storage group and destroying the RAID Group.

It may be wise to engage your local EMC service person or contact EMC support to validate and possibility to help you out.

Thanks,
Sandip

if you want to figure out which file system is using which disks the /nas/tools/.whereisfs is pretty useful

If this is an integrated Celerra where the storage has been configured using setup_clariion there is also a way to print out the current RAID and disk layout

We have one Raid 5 (8+1) that is on the same bus/encl (1_1_6 - 1_1_14) as a Raid 5 (4+1), which is on 1_1_0 - 1_1_4, with a hotspare taking up the last disk on that bus/encl.

Would that be a good candidate to break down and recreate another Raid 5 (4+1)?

Thanks again for all the good information.

BTW, the other Raid 5 (8+1) we have is on the bus/encl 0_0 with the vault drives.

Hi,

welcome to the forum.

clar_archive is a system pool which extends automatically when it finds new elibable LUNs

if after pressing rescan you dont see any new dvol's (LUNs) on the Celerra you most likely have a made a mistake when creating them.

It would be helpful to know which system you have and how you created the LUNs

Rainer

Cool info , sandip , I am going thru GUI , is there any way we can extend Clar_Archive pool . I tried to rescan but it did not get any usable luns or raid group. What we can do in these situation.

One of my colleague told me that there are plenty LUNs free in Clariion but I don't know how to add them to storage pool , Please let me know your thoughts