Single direct attached fiber for 7 vm's

[jpmfresno]

I'm currently running 7 vm's with a single direct attached fiber cable going from server to storage (the other server and hba's on order) and I have sluggish performance in my Win2k3 virtual servers. I notice that when I copy data from storage to esx server's local drives its much faster than when I copy esx server to storage. I expected it to be the other way around. When I copy vm to vm (both on the fiber attached storage) its slow.

Is it possible that the single fiber connection is a bottleneck?

Any suggestions?

Thanks

[mpverr]

It certainly could be.

What kind of San? Example - EVA are active passive so there is "load balancing on the paths". Are you seeing all VM have slugging performance? What do your monitoring tools say about the storage processor load?

[jpmfresno]

This SAN is an IBM DS3400 with 12 SAS 15k drives attached to a single internal contorller that goes fiber (single) to the HBA in the server. I have the second server, controller and hba's coming but they are weeks out. I'm not sure how to view the stroage processor load at this point. Disk I/O appears low when viewing via the VI performance page and esxtop.

[mpverr]

If you can - for starters, turn off all VM and try with just one VM. monitor closely the /var/log/vmkernal and vmkwarning files.

[lmontan]

How did you solve the performance problem. I'm now in the same scenario with poor disk performance:

One ESX 3.x Server direct attached to a DS3400 single-controler.

The writes to the SAN disks (DS3400 LUNs) are very slow, even if I write directly from de VMWare Service Console, e.g. with the dd command. Is there any HBA parameter (in the HBA BIOS or in the VMWare driver) or something similar

Your help would be very appreciated.

[Texiwill]

Hello,

even if I write directly from de VMWare Service Console, e.g. with the dd command.

All writes from the SC are throttled and expected to be extremely slow. This is by design.

Is there any HBA parameter (in the HBA BIOS or in the VMWare driver) or something similar

Review your /var/log/vmkernel log file for 'SCSI' issues. Until we know the errors that are appearing it would be difficult to determine any possible solution. I use a DS3400 with no issues at all.

Best regards,

Edward L. Haletky

VMware Communities User Moderator

====

Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.

Blue Gears and SearchVMware Pro Blogs: http://www.astroarch.com/wiki/index.php/Blog_Roll

Top Virtualization Security Links: http://www.astroarch.com/wiki/index.php/Top_Virtualization_Security_Links

[lmontan]

I attach the log files in /var/log/vmkernel. There are a lot of SCSI Warning errors, but I don't know what it means

Tks in advance,

Joan

[mike_laspina]

Hi,

Looks like device 7 is the source of your issue.

Dec 18 19:37:42 inox9002 vmkernel: VMWARE SCSI Id: Id for vmhba2:C0:T0:L7 0x53 0x58 0x37 0x33 0x38 0x30 0x30 0x31 0x36 0x38 0x20 0x20 0x20 0x20 0x20 0x20 0x31 0x37 0x32 0x36 0x2d 0x34

Dec 18 19:37:42 inox9002 vmkernel: 2:06:54:55.857 cpu5:1039)WARNING: SCSI: 2252: The physical media represented by vmhba2:0:7 has changed. The device cannot be re-synchronized with the system. This is a critical error.

Can you take LUN 7 offline?

[juchestyle]

Based on what you are talking about, it seems your storage is the issue. Fiber is very fast, actually the fastest medium we have right?

How many vms are sharing that one fiber line? Turn all of them off except one and see if it makes a difference when you transfer files. I bet it doesn't, which would mean it isn't the fiber it is another aspect.

Matthew

Kaizen!