3ware 16 port RAID-5 with 300GB Western Digital

misc technology

I am running a 9505 16 port 3ware card with 12 Western Digital 300 GB drives. In /var/log/messages I found

Nov 4 10:48:53 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x23268EDA.
Nov 4 10:50:19 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x21D88B0F.
Nov 4 10:51:27 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x1E71DB4A.
Nov 4 10:51:33 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x23289645.
Nov 4 11:02:03 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x2111C31E.
Nov 4 11:06:00 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x2219A8F3.
Nov 4 11:08:52 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x1437A499.
Nov 4 11:09:05 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x23455701.
Nov 4 11:09:10 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x23455749.
Nov 4 11:10:02 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x241F28D3.
Nov 4 11:11:54 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x20B6CCC9.
Nov 4 11:12:13 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x22277DFD.
Nov 4 11:12:13 her2 kernel: 3w-9xxx: scsi0: AEN: WARNING (0x04:0x0023): Sector repair completed:port=11, LBA=0x22277D80.
Nov 4 11:12:13 her2 kernel: 3w-9xxx: scsi0: AEN: ERROR (0x04:0x0002): Degraded unit:unit=0, port=11.

in tw_cli the drive on port 11 got reported as failed:

p11 DEVICE-ERROR u0 298.09 GB 625142448 WD-WCAPD3118453

I tried to test the drive via

/usr/sbin/smartctl -t long -d 3ware,11 /dev/twa0
/usr/sbin/smartctl -t offline -d 3ware,11 /dev/twa0
/usr/sbin/smartctl -t conveyance -d 3ware,11 /dev/twa0
/usr/sbin/smartctl -t short -d 3ware,11 /dev/twa0

But

/usr/sbin/smartctl -a -d 3ware,11 /dev/twa0

did not show many good signs:

SMART Self-test log structure revision number 1
Num Test_Description Status Remaining LifeTime(hours) LBA_of_first_error
# 1 Short offline Completed: read failure 90% 6702 396331771
# 2 Conveyance offline Completed: read failure 90% 6702 396331771
# 3 Extended offline Completed: read failure 90% 6695 396331771
# 4 Extended offline Completed: read failure 90% 6695 396331771
# 5 Extended offline Completed: read failure 90% 6694 396331773
# 6 Extended offline Completed: read failure 70% 6684 99434677

Values like Multi_Zone_Error_Rate and Offline_Uncorrectable as well as Current_Pending_Sector promised nothing good.

In tw_cli I then removed the drive in question from the unit:

maint remove c0 p11

The smarctl tests still failed right away. I rescaned the drives in tw_cli:

maint rescan c0

The failed drive was found, and soon after the 3ware controller grabbed it automatically and started the rebuild.

It did so sucessfully. The Current_Pending_Sector value decreased back to 0, and the drive array seems to be functioning
normal right now.

During one of those pesky spurious rebuilds happening on both 9550SX-16ML controllers that I am aware
of the drive failed again. This time with an ECC-ERROR . Not enough of a failure it seems that the rebuild would have
failed. A

maint rescan c0

in tw_cli after the rebuild had finished cleared this error. It’s noteworthy that the spurious rebuild performance came to a grinding
slowdown after the ECC-ERROR.

When I replaced the failed drive things went back to normal and the system has been fine ever since.

war

art internet politics technology

not your usual link or idea

War really sucks. It happens if people feel they can gain from it.

links

internet

calendar

music

rice

nyc

xray

portal

Looking and finding

art BlogsNow

Matt Stuart’s pictures are really really nice. He writes that he needs optimism for his work. I like that. I also like that much of his commissioned work is almost as good as the pictures that I would be believe are real found ones.

more predictable, yet still kinda nice are these images from the seventies

Both links I found this morning on BlogsNow. I spend some time yesterday with it. Brought the specific views back, that I had droped a while back. I still like it.

nasa

history

your tax dollars hard at work

Well, the author is probably the cousin of some senator. Better he clutters the WWW (love to use that term here and there) than if it would try to do something that actually would be needed.

OS X 10.4.11 and novatel sprint EVDO

Apple technology

Entering the country again I found that my Sprint / EVDO was not working anymore. Of course I had been stupid enough to upgrade to 10.4.11 while I was gone. Overall I am not even sure what actually caused the problem. Sprint was trying to be helpful, but had no clue. I spent fifty minutes on the phone with them. Actually only 5 minutes o those I was on hold. Somehow they changed their account system, their server kicks out a 500 error, their system assumes a cellphone to send a text message with a confirmation to. etc.

I got my EVDO going with the help of this helpful post. These were my steps:

1) System Preferences -> Network
2) Show: Network Port Configurations
3) deleted Novatel Wireless CDMA ( I had 2 !)
4) inserted Novatell U720 Modem and got “A new network port has been detected” message
5) Seleting this in “Show: Network Status” and in PPP tab enter phone number #777
6) in PPP Options … enable “Send PPP echo packets” and “use TCP header compression”
7) under Modem tab select “WWAN Support” as a Modem
8) enable ‘show modem status in menu bar’

This did do the trick. Reads longer and more complicated than it actually is.

google mail: nice new feature

google

Google mail seemed to have gotten a facelift a few weeks ago. But it also got more functionalities. Of the nice kind: I was working on an email exchange with a couple of people just now. While typing along there was all of a sudden a yellow message in the lower right corner. It informed me that somebody had added a message to the thread I was writing in. It’s this kind of helpful idea and feature that makes gmail the best mail app around these days. Very impressive.

No news in the West

history media

Sony supports Divx on their PS3

When DVD came out there was a format war as well. Between what we know as DVD and something also called Divx. Not related to the format that Sony now supports. Or maybe just related in that they represent the opposite corners of the media/business spectrum: The old Divx was a system that would only play discs after they had been enabled via the phone (it’s that old). You would buy cheaper movies (around 5-8 US) and could watch them for 48 hours after the purchase. After that you would need to pay if you like to watch them again. Studois like Dreamworks held out of for years betting on this format. It folded. But it was a reality. Hard to believe these days.

According to Sony’s CEO the format war between HD-DVD and Bluray got into a stalemate. Great thing for a format to go with this concept into the shopping season. Some big shopping mall sold HD-DVD players for 100 US$. If there would be HD-DVD and Bluray players for a 100$ each then the format war would be over. Peaceful coexistense. Not the margins that the hardware vendors had hoped for. But the luckier ones of them, if not most, make flat screen panels anyway. Those run well above one thousand dollars. And each HD-DVD and bluray playe would be an argument for a 1920×1080 panel. If those companies would be able to explain this to their clients.

Having two players may sound annoying, but most people have heaps of devices around their TVs anyway: Game consoles, Set Top Boxes, Tivo & Co, DVD players, VHS deck, another two devices might not matter. Of course all this crap looks pretty hideous once piled up. And then there is the remote control debacle. And connecting the stuff is an interesting challenge.

Americans spend more than a billion hours a day in front of their entertainment centers. And -boy- do they spend money on hard- and software. Yet, the average ‘media temple’ they pay service to each day is not much more than pile of crap.

So what about PS3 and the new-Divx? Hm. I have no idea. All I know is that the PS2 will outsell the PS3 this Christmas again, and I doubt that will change next year.

money

economy politics

According to this article China holds 1.4336 trln US$.
Big number. Each Chinese seems to make 1500 dollars a year. There are 1,321,851,888 Chinese. Which would mean that all Chinese would not need to work for 8.6 months if they would just spend the Dollars that the chinese reserve holds.

Or they could finance a seven year Iraq war. Out of the cash they have sitting around. And with the kind of spending the US military got used to. Something tells me that the red army has a different pricelist then the military in the US.

So, it seems that having stuff made over there for years somehow gave them allot of money. And therefor power.

not a hosting company to go with

history internet technology

It seems as if the some of the internet is run by amateurs. Of course an article in ‘pcworld’ is not really -ever- a testament of the truth. But, I could imagine that some hosting companies are, let’s say, scary. Problem in general is that you interface with the sales efforts not with the part that does the actual product. So it might have been that “NaviSite” had a brilliant sales team, which brought it and kept it in business. But in the end, it’s always engineering that makes or breaks things. Engineering without any marketing does not work, but also does not cause any harm. No so if there is good marketing and bad engineering. That is a common and painful combo.