tonyswu's picture

How to Patch In to Fiber Switch


Supposedly you have 2 fiber switches, and your servers / storage devices all have 2 fiber connectors, do you connect 1 to each switch, or do you connect both to 1 switch and rely on the switch uplink?

Just being curious.

AnMAX's picture

NEED HELP Xsan volume is not mounted (strange problem)

Hi Xsan guru :)
I ask the help in solving the volume mount problem. (advance I am sorry for my english)
All began with the fact that the MDC is rebooted and on the two LUNs were gone XSAN label. I relabeled these LUNs using commands:
cvlabel -c >label_list/code
then in the file label_list I corrected unknown disks on the label with a same names that were. Then I ran the [i]cvlabel label_list/i command. Finally I got the correct label on all drives.



  1. cvlabel -l

/dev/rdisk14 [Raidix meta_i 3365] acfs-EFI "META_I"Sectors: 3906830002. Sector Size: 512. Maximum sectors: 3906975711.
/dev/rdisk15 [Raidix QSAN_I 3365] acfs-EFI "QSAN_I"Sectors: 7662714619. Sector Size: 4096. Maximum sectors: 7662714619.
/dev/rdisk16 [Raidix meta_ii 3365] acfs-EFI "META_II"Sectors: 3906830002. Sector Size: 512. Maximum sectors: 3906975711.
/dev/rdisk17 [Raidix 2k_I 3365] acfs-EFI "2K_I"Sectors: 31255934943. Sector Size: 512. Maximum sectors: 31255934943.
/dev/rdisk18 [Raidix 2k_II 3365] acfs-EFI "2K_II"Sectors: 31255934943. Sector Size: 512. Maximum sectors: 31255934943.
/dev/rdisk19 [Raidix QSAN_II 3365] acfs-EFI "QSAN_II"Sectors: 7662714619. Sector Size: 4096. Maximum sectors: 7662714619.

The volume [2K] starts successfully.


but not mounted on the MDC and client.
I ran a volume check
sh-3.2# cvfsck -wv 2K
Checked Build disabled - default.


  1. !@$ Revision 4.2.2 Build 7443 (480.8) Branch Head
  2. !@$ Built for Darwin 12.0
  3. !@$ Created on Mon Jul 29 17:01:44 PDT 2013

Created directory /tmp/cvfsck3929a for temporary files.

Attempting to acquire arbitration block... successful.

Creating MetadataAndJournal allocation check file.
Creating Video allocation check file.
Creating Data allocation check file.

Recovering Journal Log.

Super Block information.
FS Created On : Wed Oct 2 23:59:20 2013
Inode Version : '2.7' - 4.0 big inodes + NamedStreams (0x207)
File System Status : Clean
Allocated Inodes : 4022272
Free Inodes : 16815
FL Blocks : 79
Next Inode Chunk : 0x51a67
Metadump Seqno : 0
Restore Journal Seqno : 0
Windows Security Indx Inode : 0x5
Windows Security Data Inode : 0x6
Quota Database Inode : 0x7
ID Database Inode : 0xa
Client Write Opens Inode : 0x8

Stripe Group MetadataAndJournal ( 0) 0x746ebf0 blocks.
Stripe Group Video ( 1) 0x746ffb60 blocks.
Stripe Group Data ( 2) 0xe45dfb60 blocks.

Inode block size is 1024

Building Inode Index Database 4022272 (100%).
4022272 inodes found out of 4022272 expected.

Verifying NT Security Descriptors
Found 13 NT Security Descriptors: all are good

Verifying Free List Extents.

Scanning inodes 4022272 (100%).

Sorting extent list for MetadataAndJournal pass 1/1
Updating bitmap for MetadataAndJournal extents 21815 ( 0%).
Sorting extent list for Video pass 1/1
Updating bitmap for Video extents 3724510 ( 91%).
Sorting extent list for Data pass 1/1
Updating bitmap for Data extents 4057329 (100%).

Checking for dead inodes 4022272 (100%).

Checking directories 11136 (100%).

Scanning for orphaned inodes 4022272 (100%).

Verifying link & subdir counts 4022272 (100%).

Checking free list. 4022272 (100%).
Checking pending free list.

Checking Arbitration Control Block.

Checking MetadataAndJournal allocation bit maps (100%).
Checking Video allocation bit maps (100%).
Checking Data allocation bit maps (100%).

File system '2K'. Blocks-5784860352 free-3674376793 Inodes-4022272 free-16815.

File System Check completed successfully.
check not helping :(
[code]sh-3.2# cvadmin
Xsan Administrator

Enter command(s)
For command help, enter "help" or "?".

List FSS

File System Services (* indicates service is in control of FS):
1>*2K[0] located on big.local:64844 (pid 5217)

Select FSM "2K"

Created : Wed Oct 2 23:59:20 2013
Active Connections: 0
Fs Block Size : 16K
Msg Buffer Size : 4K
Disk Devices : 5
Stripe Groups : 3
Fs Blocks : 5784860352 (86.20 TB)
Fs Blocks Free : 3665561306 (54.62 TB) (63%)

Xsanadmin (2K) > show
Show stripe groups (File System "2K")

Stripe Group 0 [MetadataAndJournal] Status:Up,MetaData,Journal,Exclusive
Total Blocks:122088432 (1.82 TB) Reserved:0 (0.00 B) Free:121753961 (1.81 TB) (99%)
MultiPath Method:Rotate
Primary Stripe [MetadataAndJournal] Read:Enabled Write:Enabled

Stripe Group 1 [Video] Status:Up
Total Blocks:1953495904 (29.11 TB) Reserved:270720 (4.13 GB) Free:129179 (1.97 GB) (0%)
MultiPath Method:Rotate
Primary Stripe [Video] Read:Enabled Write:Enabled

Stripe Group 2 [Data] Status:Up
Total Blocks:3831364448 (57.09 TB) Reserved:270720 (4.13 GB) Free:3665432127 (54.62 TB) (95%)
MultiPath Method:Rotate
Primary Stripe [Data] Read:Enabled Write:Enabled

I checked the availability of LUNs on MDC and client, there also all right.

But, unfortunately, the Volume is not mounted :(

[code]sh-3.2# xsanctl mount k2
mount command failed: Unable to mount volume `k2' (error code: 3)

[b]Please help me to figure out this situation, I will be grateful for any information./b

lucasnap's picture

_Inodelookup invalid inode [0x0]


We've just upgraded our Xsan volumes from 10.6.8 Xsan 2.2.2 to 10.8.5.
This went well, apart for one issue.

I needed to reset all the permissions (we went from OD to AD) and while propagating the permissions (and touching all the files) I get about 500 of these errors:

fsm[3962]: Xsan FSS 'XsanVolume[0]': _Inodelookup invalid inode [0x0]
fsm[3962]: Xsan FSS 'XsanVolume[0]': _Inodelookup invalid inode [0x0]
fsm[3962]: Xsan FSS 'XsanVolume[0]': _Inodelookup invalid inode [0x0]

Always the address 0x0.

I've run cvfsck -jv and cvfsck -nv but the check doesn't find anything.

Interesting is that after the restart i've propagated the permissions again and then they are gone. Until the next morning.. I wanted to be sure and run the propagate again, now the error's are back.

Anyone seen these? Google isn't very helpful.



Xsan 3: Invisible "__defragtmp" files are created after using snfsdefrag (Apple KB)

When defragmenting a file with snfsdefrag on systems running OS X Mountain Lion v10.8.4 and earlier, an invisible "__defragtmp" file may be created which cannot be deleted.
This issue may occur when the following are true:

The defragmented file's name contains precomposed Unicode characters, such as letters with diacritical marks (an accent, grave, or umlaut).
snfsdefrag is run by root.
snfsdefrag is run when the current working directory is on an Xsan volume, or snfsdefrag's recursive option (-r) is used.

Read more:

buchau's picture

XSan Permission Issue and ..

We currently moved from
Mac OSX 10.6.x MDCs using XSAN 2.2.x / Mac OSX Open Directory / XSAN 2.2.x Mac OS X 10.6.x Clients / Promise Storage ( all connected via Fiber Switch )
StorNext MDCs ( M440 Appliance ) version 4.7/ New Storage / Old Promise Storage was left intact and still being utilized / New MAC OS X 10.8.x XSAN Clients / Two new Windows 2008 Servers utilizing Stornext Clients All still connected via Fiber Switch.

All The Fiber connected Clients Mac OS X 10.8.x ( XSAN Client ) via Fiber & Two Windows 2008 R2 Server ( Stornext Client ) can see the old ( XSAN ) LUNs & the new ( StorNext ) LUNs and mount the volumes ( lets call this xsan1 and snfs1 ) I can see and access ALL the folders via the MAC OS X 10.8.x fiber clients but via Windows I see the volume but am not able to access "all" the contents. At first I saw the generic permission message indicating " You don't currently have permission to access this folder. Click continue to permanently get access to this folder"
At which point I can choose to continue, and take ownership of the folder. that works to a certain extent but I still get access denied while taking ownership of multiple folders and so on ..... I have been reading up on this issue and came across few things but figured I would come to the PROs and see what is the best way to approach this issue.
I also would like to utilize my Active Directory Domain to authenticate users to login to the MACs and also permission to the old Xsan volume and new snfs volume.
any help is greatly appreciated and I look forward to hearing from the PROs

eWhizz's picture

Expected speed? Promise Vtrack E-Class 32TB (16 x 2TB SAS)

What would the expected range of speed be from the following —

Apple Promise VTrack E-Class 32TB RAID Subsystem for MAC OS X w/ 16x 2TB SAS

We would expect to create two RAID6 LUNS and then stripe them together using Mac OS X Software RAID Striping and share the resulting volume using MetaSAN.

I know you guys are more specialising in Xsan rather than MetaSAN, but you are also the most likely group of people to know the answer to this one.

Thanks, in advance.

jetsenwang's picture

About 10.8.3 Xsan admin and 10.6.8 client

About 10.8.3 Xsan admin and 10.6.8 client

MetaData Controller : 10.8.3 Server Xsan3.0
Client:10.6.8 Xsan 2.2.2

How to connect?????

DLpres's picture

New SAN, can't add any clients (Failed to write conf. files)

This is a fresh XSAN, I started with a clean install on a wiped 10.8.5 ML Server (Xserve 2009). Have LUNs, volume, and already migrated some data to it. So far so good.

I tried for 2 hours to add (10.8.5 and 10.8.4) clients with no success. I enable XSAN in SysPrefs, XSAN Admin sees the clients easily, I authenticate, but after a while I get the message:
"Failed to write configuration files on computer"

The first 'client' was another server which I was going to promote to a secondary MDC, but never got this far. The other client was a normal Mac Pro.

I checked pings, DNS (dig), changeip -checkhostname, everything resolves just fine. OD is in place, separate private metadata network set manually with a separate subnet.

Are there any log files that'll show in more detail what's failing and where?

One thing I just thought about is that our entire facility DNS is set with A records for machines ("Solo") and CNAME for roles ("edit1"). The Add Computers window shows the CNAMEs for the machines, which looks just the way I expect it to, but is there something against having CNAMEs in an XSAN environment?

matt321's picture

10.8.5 Update and FinalCut X (10.0.9)

We are having issues since the 10.8.5 update with FinalCut X (10.0.9) randomly unmounting XSAN volumes.

Only rebooting the workstations fixes the issue.

Anyone having similar issues since updating?

DLpres's picture

Anyone got Unscale's HD Speed Test utility?


This little piece of software was touted here before as a great benchmark tool, since it generates data the same way an NLE does, giving a better simulation of video environment.

The problem is, Unscale has gone poof, their blog is gone, their website became Chinese, and it seems that there are zero copies of the software available elsewhere.

The latest version should be 1.2.2, does anyone have it?


Subscribe to Xsanity RSS