Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Ubuntu file system going into "lockdown" mode



Adding to what Jerry observed:
> In the past I have had to replace ribbon cables after encountering
> errors... I suspect that the issue with
> that system is just that it is not aging gracefully, and would do much
> better if submerged in Salem harbor :-)

Every few years I take the time to toss out old hardware that no longer has
enough oomph to handle whatever shiny new application I want to play with. 
(This year it's VirtualBox, which although fun does like to snack on RAM and
disks for lunch, dinner, midnight snack, Sunday brunch & every other chance it
gets...)

So--my point is that even some of that new hardware one buys can produce error
messages similar to those reported here.

In my case the problem showed up as disk drives dropping out of a software
RAID array that I was using to make backups.  What I noticed was this item in
the output from 'smartctl -a /dev/sdX':

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED 
WHEN_FAILED RAW_VALUE
...
199 UDMA_CRC_Error_Count    0x000a   200   200   000    Old_age   Always      
-       98

Hmm, a non-zero CRC error counter on a relatively new SATA drive?!?  Googling
for the error was not a lot of help, but at least I learned that this problem
is not the drive electronics themselves but rather the connection between the
computer and the drive.

If you look up SATA specifications at Wikipedia, you'll note that the
cable-length specification is 1 meter (for any speed 1.5/3.0/6.0 Gbit/sec).

Most SATA drives have only one jumper selection; on Seagate and Hitachi drives
like what most of mine are, it allows you to change the interface speed.  On a
hunch, I tried changing the interface speed from 3.0 to 1.5.  Voila, my
problem was resolved!  I still get a small number of errors on my setup but
drives are no longer dropping offline.

Explanation:  a couple years ago I bought an inexpensive 12-bay external
enclosure from a manufacturer called Norco.  It has 3 Infiniband connectors on
the back.  Opening up the case, I found that the drives are connected to these
connectors using ordinary consumer-grade 18-inch SATA cables.  Adding that
length to the 1-meter Infiniband cables I am using is enough beyond the SATA
specification to cause me a heap of trouble.

Lesson learned is that if you're dealing with external drive bays, you need to
watch your cable lengths or switch to (vastly) higher-cost SAS technology. 
And avoid Norco like the plague, at least if their current products are
engineered as badly (in several other ways besides the one I mention here) as
my 2008-vintage 12-bay chassis.  Note that eSATA will let you go to 2m cable
lengths; although I see eSATA ports on a lot of devices these days, I don't
see that technology going anywhere now that USB 3.0 is out (5 GBps).  That's
kinda too bad because I like the diagnostics you get with 'smartctl'--drives
attached via USB don't respond correctly to that program.

I see on Newegg there is a rival to Norco called iStarUSA which makes
inexpensive enclosures (example, the 10-bay DAGE410U20-PM-RED).  Are their
products any good?  I got spoiled by Dell and HP enclosures at work, but those
are priced beyond a consumer's budget for mere backups.

-rich







BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org