VMware disk errors on guest machine

crazyivan · 07-16-2008, 04:43 AM

Hello Crew!

The past few months I've been getting the following errors on my Debian guest machine. (Please see below, logs taken from logwatch) I experience very bad virtual disk performance when copying a lot of files. Disk access time is very slow too. I've read the issues with the LSI and BusLogic and choose to use the LSI adapter.

- Can someone explain the error messages?
- Has anybody seen these errors before?
- Might this be a known issue? (If so I have not found anything about it)

Any help is greatly appreciated.

Code:

--------------------- Kernel Begin ------------------------ 


1 Time(s):         command: Write(10): 2a 00 00 04 4d 17 00 00 18 00
1 Time(s):         command: Write(10): 2a 00 04 8c 11 57 00 00 08 00
2 Time(s): mptbase: ioc0: IOCStatus(0x004b): SCSI IOC Terminated
1 Time(s): mptscsih: ioc0: attempting task abort! (sc=c5c7b980)
1 Time(s): mptscsih: ioc0: attempting task abort! (sc=e29df980)
1 Time(s): mptscsih: ioc0: task abort: SUCCESS (sc=c5c7b980)
1 Time(s): mptscsih: ioc0: task abort: SUCCESS (sc=e29df980)
2 Time(s): sd 0:0:0:0: 

---------------------- Kernel End -------------------------

Host(openSUSE 10.3 (i586)): Linux host 2.6.22.18-0.2-default #1 SMP 2008-06-09 13:53:20 +0200 i686 i686 i386 GNU/Linux
Using VMware Server 1.0.6

Guest (Debian GNU/Linux 4.0): Linux guest 2.6.18-6-686 #1 SMP Fri Jun 6 22:22:11 UTC 2008 i686 GNU/Linux
VMware Tools 1.0.6 installed

Guest VMware configuration: servername.vmx

Code:

#!/usr/bin/vmware
config.version = "8"
virtualHW.version = "4"
scsi0.present = "TRUE"
scsi0.virtualDev = "lsilogic"
memsize = "2048"
scsi0:0.present = "TRUE"
scsi0:0.fileName = "disk3-000001.vmdk"
scsi0:0.writeThrough = "TRUE"
ide1:0.present = "TRUE"
ide1:0.fileName = "/usr/lib/vmware/isoimages/linux.iso"
ide1:0.deviceType = "cdrom-image"
floppy0.startConnected = "FALSE"
floppy0.fileName = "/dev/fd0"
Ethernet0.present = "TRUE"
displayName = "servername"
guestOS = "other26xlinux"
priority.grabbed = "normal"
priority.ungrabbed = "normal"
powerType.powerOff = "hard"
powerType.powerOn = "hard"
powerType.suspend = "hard"
powerType.reset = "hard"

floppy0.present = "FALSE"

scsi0:0.redo = ""
ide1:0.startConnected = "FALSE"
ethernet0.addressType = "generated"
uuid.location = "56 4d 55 4b 3d c2 0c 56-43 3a c1 ff af f1 19 32"
uuid.bios = "56 4d 55 4b 3d c2 0c 56-43 3a c1 ff af f1 19 32"
ethernet0.generatedAddress = "00:0c:29:f1:19:32"
ethernet0.generatedAddressOffset = "0"

autostart = "poweron"
# Disable Time Synch
tools.syncTime = "FALSE"
tools.synchronize.restore = "FALSE"
time.synchronize.resume.disk = "FALSE"
time.synchronize.continue = "FALSE"
time.synchronize.shrink = "FALSE"

checkpoint.vmState = ""

scsi0:1.present = "TRUE"
scsi0:1.fileName = "emdb01-000001.vmdk"

scsi0:1.redo = ""

numvcpus = "2"

usb.present = "TRUE"
usb.generic.autoconnect = "FALSE"

ide0:0.present = "FALSE"
ide0:0.fileName = "emdb IDE-000001.vmdk"
workingDir = "."
monitor_control.log_vmsample = "TRUE"

ide0:0.redo = ""

zeno0771 · 08-03-2008, 12:04 AM

Is the guest on the same drive as the host? VMware guest performance suffers mightily if the host, VMware, and/or the guest are being read from the same drive, and VMware (the Workstation version anyway) puts VMs in a folder in the user's home directory by default. Even a separate partition doesn't help matters in my experience. Are you using actual SCSI drives? I've only seen one instance where SCSI was recommended over IDE for a guest install; SATA drives work just fine in IDE and it saves you the trouble of choosing between SCSI adapters entirely.

jiml8 · 08-03-2008, 12:12 AM

You are getting hard errors that may indicate a hardware issue. I have not seen this out of the VMWare system before, so I don't know how the VMWare SCSI emulators will handle a hard error, but if I were you I would shut down the guest and start running host tools on the hard drive that hosts the guest.

The indicated errors could mean a hard drive failure is underway, but they also could indicate that you have a controller failing.

You need to find out whether either of these is the case. If you have a controller failing, then you probably have other indications of I/O errors in your hard drives; check what is appearing on your default console and what is in /var/log/messages

crazyivan · 09-18-2008, 05:16 AM

The host OS and the guest VMDK reside on the same software SATA mirror on the physical host. I can see why this could give performance issues but not why my logs fill up with errors.

I've done checks on my host OS and it's drives. They seem fine. All drives still have a throughput of 75-80 Mb/s

These are the messages I get when I have a lot of drive activity. A few per hour, and during such an error the whole system becomes unavailable and slow as molasses.

Code:

Sep 17 11:58:06 smurfer3nieuw kernel: mptscsih: ioc0: attempting task abort! (sc=deaf6ac0)
Sep 17 11:58:06 smurfer3nieuw kernel: sd 0:0:0:0: 
Sep 17 11:58:06 smurfer3nieuw kernel:         command: Write(10): 2a 00 04 77 4c 77 00 00 10 00
Sep 17 11:58:06 smurfer3nieuw kernel: mptbase: ioc0: IOCStatus(0x004b): SCSI IOC Terminated
Sep 17 11:58:06 smurfer3nieuw kernel: mptscsih: ioc0: task abort: SUCCESS (sc=deaf6ac0)

Any suggestions?