Archive for the ‘tech’ Category

HFS+ file system corruption on Tiger (and maybe leopard) Intel x86 Mac Pro

July 11, 2008

Hello again,

This is one of those issues that has very little covereage on the web so hopefully this will help someone.

The majority of our workstations are Mac Pro desktops and we have quite a large number of them.  Due to this we tend to see problems that most people would not see with just a few workstations.  One of the most irritating problems these days seems to be exclusive to the newer Intel x86 Mac Pro workstations running TIger.  We have not run Leopard long enough to determine if it also has this issue.

The root volume will start to show file system corruption over a short period of time, but HFS and/or OS X will not reporting it to the user or in any of the log files and continue to make the corruption worse.  If the corruption is caught early on a simple Single User mode fsck will fix the issue.  If it not caught for some time then Disk Warrior is needed to repair the volume.  And if the issue is left alone too long even Disk Warrior will not help and a reformat of that partition (and re-install, being the root partition) is necessary.

These system are all on UPS Battery backup power so they do not shutdown uncleanly.  The systems are centrally managed and patched and updated regularly.  Only the Intel x86 versions show these symptoms.  PowerPC versions of our Mac Pros do not show any of this corruption.  The only reboots on these systems happen for System Updates so it is possible that System Update is rebooting the system before the file system is fully synced after the updates.

In order to detect this problem we routinely run a Verify with Disk Utility.  This can be done on the commandline with:

diskutil verifyVolume /

This command can be integrated into a cronjob which will email the administrator if a problem is detected.

Hopefully Leopard on x86 Mac Pro does not start showing these same symptoms, but time will tell.  Another post will follow if Leopard does show the same symptoms.

Hello world!

July 11, 2008

Hi,

The goal of this blog is an attempt to document some issues that have been observed with various operating systems, software, hardware and networking.  Often a blog is the only Google-able source for information about a particular quirk with an operating system or piece of software.  Hopefully this blog will contribute.

The setting is a fairly large NASA funded space mission with a complex networking and computing environment.  The issues that will be discussed will be technical in nature and no mission specific information will be discussed.

Hopefully this information will be useful to others having the same problems since quite a few of the (soon to be posted) issues are not easily Google-able.

Technologies used:

Sun hardware, OpenSolaris, Solaris 10, SuSE Linux, OS X Tiger, OS X Leopard, VPN, IPSec, OpenVPN, ZFS, AVS, NAS, Netapp, Mac Pro, MacBook Pro, HFS, Cisco, Fibre Channel, Comstar, XMPP, DNS, LDAP, JAVA.