HPE EVA Storage

severe performance problems on the brocade fabric

 
ausiee
Occasional Advisor

severe performance problems on the brocade fabric

Hello All,

We have EVA 8000, 2 controllers, 8 enclosures, 84 disks, 20 FATA and 64 HPFC.

Dual 4GB fabric brocade switches.

4 x C7000 blade server enclosures connecting into the fabric by 2 fibre leads, one into each fabric switch.

Each fabric switch is setup as it's own zone, and the disks are grouped into FATA disks, and HPFC disks, all VRAID5.

We had a backup device connected into the fabric on one of the fibre switches, and created a zone for the backup device and connecting blade server.

We had problems with the controller on the backup device that caused severe performance problems on the fabric, to the extent that the only way to resolve the issues was to shut down the port that the backup device connected into on the fabric switch. Even removing the backup zone did not fix the issues.

Any ideas?
5 REPLIES 5
Eemans Dany
Honored Contributor

Re: severe performance problems on the brocade fabric

Hi,

Not sure i will help.
Create a separated zone for each server that needs to see the drive.

Ex
Zone 1 = Server1 and Tapedevice1 / tapedevice 2
Zone 3 = Server2 and Tapedevice1 / tapedevice 2

Or create the zone by drive and server

and so one.

If you have degrade performance you can eleminate zone by zone to find the server / Device causing the problems.

Also try not to mix windows - Linux - unix - other together in the same zone.

Dany

Dany
ausiee
Occasional Advisor

Re: severe performance problems on the brocade fabric

Thanks mate for the advises.

Just to update;

When the backup device is attached it causes performance issues. Servers lose attached san disks and some servers reboot.

Do you think it is some thing to do with Zonning and mix-up of OS?

Appreciate your response.
Kites BM
Regular Advisor

Re: severe performance problems on the brocade fabric

>>>>>When the backup device is attached it causes performance issues. Servers lose attached san disks and some servers reboot

What speed are the backup devices running at?You need to match the speed of backup devices to the speed on port of the switch.

May be this works :)

Kites
Eemans Dany
Honored Contributor

Re: severe performance problems on the brocade fabric

Hi
"
When the backup device is attached it causes performance issues. Servers lose attached san disks and some servers reboot.

Do you think it is some thing to do with Zonning and mix-up of OS?
"

If they are all in one zone, this can happen. When the zones are separated, the tapedevice cannot send trash data to the servers and there disks.
This in the understanding the tapedevice is the problem and not another server accessing the tapedrive.

Dany
ausiee
Occasional Advisor

Re: severe performance problems on the brocade fabric

Now, I have some queries;

1 - The previous backup zone is deleted by the on-site IT staff. I have created a new zone on the brocade switch [not the BLADE SYSTEM 4GB Switch] includes MSL6000 PWWN and blade's mezzanine card PWWN. [Is it enough for the devices to see each other or should I present the BROCADE BLADESYSTEM 4/12 SAN SWITCH to this zone as well?]

2 - Previously, in the backup zone the following devices were included;

- MSL6000 PWWN
- blade's mezzanine card PWWN
- BROCADE BLADESYSTEM 4/12 SAN SWITCH PWWN
- Switch 1 WWN
- Switch 2 WWN

Doesn't look fair to me - your thoughts please?

Your earliest response is highly appreciated.