HP 3 par storeserv 7200 volume not accessible

[hedijaiet]

hi;
we have a hp 3 par store serv 7200 dual node with 8 disks 300 Gb 15K, after the initial and default configuration we have 2 cpg one on raid 1 and the other on raid 6 .
we create a new volume from the default cpg and we mapped it to one host.
when we tray to write on the volume we have an error message : we dont have a disk space to write on the volume

[skumflum]

Hi,
Why raid1 and raid6? For FC disk Raid5 is the recommended Raid-level in 90% of the time.

What kind of host? Windows?
Where do you get this error message? On the host or in the 3par management?

[hedijaiet]

hi;
when we lanched the first installation for the 3 par we found by default 2 cpg raid 1 and raid 6 . we dont create any cpg manualy and the sytem is installed on the raid 1 cpg .
my host is Vmware ESXI 5.1
we get the error code on the host.

[skumflum]

Create a Raid5 CPG and don't use the other ones.

I would strongly recommend that you read AND re-read the “HP 3PAR StoreServ Storage and VMware vSPhere 5 best practices”

http://www.vmware.com/resources/techresources/10102

[Richard Siemers]

I second that suggestion. The best practice/implementation guides for specific OSes (ESX) are really good.

8x300 disks is not much room for host data after you subtract spares, system overhead, raid overhead.

Open a ssh connection to the 3par and run "showpd -space" command and paste the result please.

[hedijaiet]

hi;
thank you for your help;
so the 8 first disks on 3 par is used for the system internal data and i cannot use them as volume on my hosts ?

[Richard Siemers]

so the 8 first disks on 3 par is used for the system internal data and i cannot use them as volume on my hosts ?

Not true. System space, spare space will be distributed across all the disks installed on the system , whatever space is left over will be usable for host volumes.

What does "showpd -space" tell you? Here is an example of what a 16 drive system looks like:

Code: Select all

ESFWSS7200-1 cli% showpd -space
                        ----------------------(MB)-----------------------
Id CagePos Type -State-     Size   Volume   Spare     Free Unavail Failed
 0 1:0:0   NL   normal   2772992  1219584  347136  1206272       0      0
 1 1:1:0   NL   normal   2772992  1221632  347136  1204224       0      0
 2 1:2:0   NL   normal   2772992  1221632  347136  1204224       0      0
 3 1:4:0   NL   normal   2772992  1221632  347136  1204224       0      0
 4 1:5:0   NL   normal   2772992  1221632  347136  1204224       0      0
 5 1:6:0   NL   normal   2772992  1221632  347136  1204224       0      0
 6 1:8:0   NL   normal   2772992  1219584  347136  1206272       0      0
 7 1:9:0   NL   normal   2772992  1221632  347136  1204224       0      0
 8 1:10:0  NL   normal   2772992  1221632  346112  1205248       0      0
 9 1:12:0  NL   normal   2772992  1219584  346112  1207296       0      0
10 1:13:0  NL   normal   2772992  1220608  346112  1206272       0      0
11 1:14:0  NL   normal   2772992  1222656  346112  1204224       0      0
12 1:16:0  NL   normal   2772992  1221632  346112  1205248       0      0
13 1:17:0  NL   normal   2772992  1221632  346112  1205248       0      0
14 1:20:0  NL   normal   2772992  1219584  346112  1207296       0      0
15 1:21:0  NL   normal   2772992  1220608  346112  1206272       0      0
-------------------------------------------------------------------------
16 total                44367872 19536896 5545984 19284992       0      0

[hedijaiet]

thank you for your replay ;
see attached the configuration of my defaults CPGs
when i have exported my volume 3PAR-test to the HOST i have an error message: u cannot write on this volume.

[Richard Siemers]

copysize.JPG (34.69 KiB) Viewed 28353 times

Your problem is something is taking up all that copy space in your R6 cpg... find out what it is and perhaps delete it if that is the correct option for your circumstance... then compact the CPG.

[Cleanur]

Minimum disks required for raid 6 is 12 disks, this is because the system is symmetric active active, as such each controller owns half the disks.

Minimum raid 6 stripe is 4+2 which means six disks are required per controller. I have no idea how you got to this configuration, but based on 8 disks you can only do raid 1 or raid 5 (2+1) or (3+1). It looks like the CPG can't grow as directed to accommodate writes as there aren't enough disks to complete the stripes required for the underlying LD's.

I would delete the Raid 6 CPG and create a new one on raid 5 (3+1).