XSAN almost working - what do I need to kick?

Question

Level 4

1,240 points

XSAN almost working - what do I need to kick?

Here is the long and short of it all

3 - G4's Dual 1.42
7 - G5's Dual 2.5

XSAN = 1.2
OSX = 10.4.5

IP addresses of
169.254.x.x
mask of 255.255.0.0

All connected via fiber channel switch to XRAID. Internal GB Ethernet connects them all to MD controllers. Everyone can ping everyone else, so e-net is working fine.

The G4's are working and using the RAID/XSAN. The G5's cannot mount the volume and are not seen in the XSAN Admin app. If I bring up Apple System Profiler and click on the Fiber Channel, it sees both volumes just fine. If I go to /Volumes, there is nothing there (I figured there wouldn't be since I have to mount via XSAN Admin).

Below is a clip from the system.log file on the G5. I am at my ropes end trying to get this beast up and running - It has become the bane of my existance. Any suggestions would be appreciated (execpt to reboot the whole system, this has been done many a time already with no positive results).

Thank you.
Peter

Below is the clip from the log file on the G5

Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: \n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: File System Server:\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Server Revision 2.6.3 Build 27.3\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Built for Darwin 8.0\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Created on Thu Nov 3 16:45:37 PST 2005\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: \n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: \n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: File System Client:\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Client Revision 2.6.3 Build 27.3\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Built for Darwin 8.0\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Created on Thu Nov 3 16:48:45 PST 2005\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 SNFS: Built in /SourceCache/XsanFS/XsanFS-263.27.3\n
Apr 28 12:58:55 admin2s-power-mac-g5-2 sudo: root : TTY=unknown ; PWD=/Library/Filesystems/Xsan/debug ; USER=root ; COMMAND=/sbin/kextload -v -s /Library/Filesystems/Xsan/debug /System/Library/Extensions/acfs.kext
Apr 28 12:58:56 admin2s-power-mac-g5-2 kernel[0]: Xsan Client Revision 2.6.3 Build 27.3 Built for Darwin 8.0 Created on Thu Nov 3 16:48:45 PST 2005
Apr 28 12:58:56 admin2s-power-mac-g5-2 sudo: root : TTY=unknown ; PWD=/Library/Filesystems/Xsan/debug ; USER=root ; COMMAND=/sbin/kextload -v -s /Library/Filesystems/Xsan/debug /System/Library/Extensions/acfsctl.kext
Apr 28 12:58:58 admin2s-power-mac-g5-2 servermgrd: xsan: [42] main: Waited 65 secs for fsmpm to start (now running)
Apr 28 12:58:58 admin2s-power-mac-g5-2 fsmpm[251]: PortMapper: No CVFS Disk Volumes are accessible.\n
Apr 28 12:58:59 admin2s-power-mac-g5-2 servermgrd: xsan: [42] Done waiting for fsmpm to start
Apr 28 12:58:59 admin2s-power-mac-g5-2 fsmpm[251]: PortMapper: Local FSD client is registered.\n
Apr 28 12:59:01 admin2s-power-mac-g5-2 servermgrd: xsan: [42/30F800] ERROR: get localproperties: Did not save reply because no disk or status=0/0
Apr 28 12:59:03 admin2s-power-mac-g5-2 servermgrd: xsan: [42/32F1E0] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/324DF0] ERROR: is addressreachable(169.254.136.84): Address is not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/324DF0] ERROR: most likely_reachable_nic_forip(169.254.136.84): Host is unreachable and does not have a proxied session ID. Cannot determine other NICs without it. Falling back to 169.254.136.84 which is probably not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/3233E0] ERROR: is addressreachable(169.254.58.27): Address is not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/3233E0] ERROR: most likely_reachable_nic_forip(169.254.58.27): Host is unreachable and does not have a proxied session ID. Cannot determine other NICs without it. Falling back to 169.254.58.27 which is probably not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/326D90] ERROR: is addressreachable(169.254.38.121): Address is not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/326D90] ERROR: most likely_reachable_nic_forip(169.254.38.121): Host is unreachable and does not have a proxied session ID. Cannot determine other NICs without it. Falling back to 169.254.38.121 which is probably not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327840] ERROR: is addressreachable(169.254.204.190): Address is not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327840] ERROR: most likely_reachable_nic_forip(169.254.204.190): Host is unreachable and does not have a proxied session ID. Cannot determine other NICs without it. Falling back to 169.254.204.190 which is probably not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327DB0] ERROR: is addressreachable(169.254.117.134): Address is not reachable
Apr 28 12:59:04 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327DB0] ERROR: most likely_reachable_nic_forip(169.254.117.134): Host is unreachable and does not have a proxied session ID. Cannot determine other NICs without it. Falling back to 169.254.117.134 which is probably not reachable
Apr 28 12:59:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/324DF0] ERROR: get remoteproperties: Could not reach 169.254.136.84:311
Apr 28 12:59:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/3233E0] ERROR: get remoteproperties: Could not reach 169.254.58.27:311
Apr 28 12:59:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/326D90] ERROR: get remoteproperties: Could not reach 169.254.38.121:311
Apr 28 12:59:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327840] ERROR: get remoteproperties: Could not reach 169.254.204.190:311
Apr 28 12:59:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/327DB0] ERROR: get remoteproperties: Could not reach 169.254.117.134:311
Apr 28 12:59:21 admin2s-power-mac-g5-2 servermgrd: xsan: [42/32C0F0] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 12:59:28 admin2s-power-mac-g5-2 servermgrd: xsan: [42/324770] ERROR: -[XsanAutomounter threadedAutomount:]: Error automounting 'JourMetaVol' (CANNOT MOUNTERROR)
Apr 28 13:00:27 admin2s-power-mac-g5-2 servermgrd: xsan: [42/33BD50] ERROR: -[XsanAutomounter threadedAutomount:]: Error automounting 'JourMetaVol' (CANNOT MOUNTERROR)
Apr 28 13:03:21 admin2s-power-mac-g5-2 servermgrd: xsan: [42/3347E0] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 13:03:48 admin2s-power-mac-g5-2 servermgrd: xsan: [42/32BFA0] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 13:03:51 admin2s-power-mac-g5-2 servermgrd: xsan: [42/321070] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 13:03:53 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 8 (Loop State Change) for SCSI Domain = 0
Apr 28 13:03:53 admin2s-power-mac-g5-2 kernel[0]: FusionFC: Loop Initialization Packet for SCSI Domain = 0.
Apr 28 13:03:53 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 5 (External Bus Reset) for SCSI Domain = 0
Apr 28 13:03:53 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: External Bus Reset for SCSI Domain = 0
Apr 28 13:03:53 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 9 (Logout) for SCSI Domain = 0
Apr 28 13:03:55 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 8 (Loop State Change) for SCSI Domain = 0
Apr 28 13:03:55 admin2s-power-mac-g5-2 kernel[0]: FusionFC: Loop Initialization Packet for SCSI Domain = 0.
Apr 28 13:03:55 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 5 (External Bus Reset) for SCSI Domain = 0
Apr 28 13:03:55 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: External Bus Reset for SCSI Domain = 0
Apr 28 13:03:57 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 9 (Logout) for SCSI Domain = 0
Apr 28 13:03:57 admin2s-power-mac-g5-2 kernel[0]: FusionMPT: Notification = 6 (Rescan) for SCSI Domain = 0
Apr 28 13:05:19 admin2s-power-mac-g5-2 servermgrd: xsan: [42/3217A0] ERROR: +[XsanConfiguration threadedSynchronize:]: localhost (169.254.35.52) sees 0 disks
Apr 28 13:18:22 admin2s-power-mac-g5-2 /System/Library/CoreServices/loginwindow.app/Contents/MacOS/loginwindow: Login Window Application Started
Apr 28 13:18:22 admin2s-power-mac-g5-2 loginwindow[394]: Login Window Started Security Agent

AlBook 12 / Dual2.0G5 / iMac 20" IntelDuoCore, Mac OS X (10.4.4)

Posted on Apr 28, 2006 11:15 AM

Reply

Answer 1

Alex Geis

Level 2

462 points

Apr 28, 2006 8:31 PM in response to mrpetey

Looking at your host names in the log and that you're using a 169 IP set, I'm guessing you don't have DNS setup? From experience with older versions of xsan (1.1 and earlier), not having DNS setup caused serious issues, and I think it's still required by Apple if you want your xsan network to run smoothly.

Do any of your machines run OS X Server, and if not, you might want to take the route to manually setup tables for your hosts on each machine.

We have 5 clients on our end, and we have one central xserve handling DNS and MDC, and one of our clients doubling as a secondary DNS.. .the CPU load has been suprisingly low across the board.

Reply

Answer 2

sapridyne

Level 1

144 points

Apr 28, 2006 8:59 PM in response to Alex Geis

Actually, DNS is not necessary or required for Xsan to work properly. And as for the IP configuration, I HIGHLY suggest not using a DHCP (i.e. a 169.254) address -- they tend to change across reboots, which really confuses Xsan (I believe this is causing the IP errors in the log you posted).

All you need in Network Preferences is the IP address (manually entered) and the subnet mask. No DNS, no router/gateway. Xsan doesn't use/need name resolution so DNS entries aren't necessary. You're also not routing the network information across subnets, so there's no need for router/gateway information.

On a side note, it's probably not a real good idea to have your MDC doing DNS. If it's working for you, great, but it's not best practices and I suggest moving your DNS onto another machine.

Sapridyne

Reply

Answer 3

May 1, 2006 1:06 AM in response to mrpetey

Hi mrpetey,
It looks like the trouble is in your ethernet, and not in your fibre.
You should check this though, from the command prompt type cvlabel -l
This should show the lun's from which the volume is build.

If this is correct, configure the ethernet part.
You can also look at cvlog (/Library/Filesystems/Xsan/data/$VOLUMENAME/log/cvlog)

HTH
Donald

Reply

Answer 4

mrpetey Author

Level 4

1,240 points

May 2, 2006 12:02 PM in response to Donald Kok

Well, here it goes. After some more tinkering this afternoon, I have some more info.

All 169.254.x.x addresses are entered manually. I figured the auto-assign would screw things up. Everyone can PING everyone else, so everyone does see one another and the e-net is up and working.

Primary MDC reports back all of the RAID volumes. The G5's report back nothing with the cvlabel command. ASP does report back the volumes when I click on the Fiber Channel selection (on the G5's).

One new thing. When I have open a window for /Volumes. I see the three local volumes. An icon for the main RAID drive comes up and blinks. Slow enough you can see it, but fast enough you can't click on it.

Reply

Answer 5

May 2, 2006 11:38 PM in response to mrpetey

Hi,
I think you should focus on the fibre part of your system.
cvlabel -l should give you the lun's. Probably your fibreswitch is not configured for the G5's (are they new to the xsan?) check out zoning on the fibre switch.
Donald

Reply