EsoErik

When the first cylindrical acrylic reservoir installed in my modified, liquid cooled Shuttle workstation cracked, I blamed myself. Perhaps I overtightened the reservoir end caps, or induced excessive strain as I installed it in its cramped location, I thought.

Now, the replacement reservoir, treated ever so gingerly, has also cracked:

This second, identical failure set me to researching the problem more diligently. The cause was not difficult to identify: Water Wetter is known in the PC water cooling community to cause crack development in acrylic/Plexiglas components. So, I am purchasing a third reservoir for this system, and I'm using distilled water + a few drops of iodine as coolant. I'll add surfactant and corrosion inhibitors once I am sure I have found ones that will not cause my cooling loop to shatter!

Labels: water cooling

This morning, I tripped a breaker that happens to include my workstation. After distributing load to other circuits and powering up my workstation again, I was confronted by an ugly surprise in my dmesg:

[ 2336.676714] ata8: SError: { CommWake DevExch }
[ 2336.676720] ata8: hard resetting link
[ 2340.032036] ata8: SATA link down (SStatus 0 SControl 300)
[ 2340.048016] ata8: EH complete
[ 2344.455985] ata8: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
[ 2344.455985] ata8: irq_stat 0x00000040, connection status changed
[ 2344.455985] ata8: SError: { CommWake DevExch }
[ 2344.455985] ata8: hard resetting link
[ 2354.476044] ata8: softreset failed (device not ready)
[ 2354.476047] ata8: hard resetting link
[ 2364.528028] ata8: softreset failed (device not ready)
[ 2364.528032] ata8: hard resetting link
[ 2375.484019] ata8: link is slow to respond, please be patient (ready=0)
[ 2399.580028] ata8: softreset failed (device not ready)
[ 2399.580033] ata8: limiting SATA link speed to 1.5 Gbps
[ 2399.580035] ata8: hard resetting link
[ 2404.780022] ata8: softreset failed (device not ready)
[ 2404.780025] ata8: reset failed, giving up
[ 2404.780031] ata8: EH complete

I should have seen something more like this:

[ 4704.104020] ata7: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 4704.104251] ata7.00: ATA-9: C300-CTFDDAC256MAG, 0006, max UDMA/100
[ 4704.104253] ata7.00: 500118192 sectors, multi 1: LBA48 NCQ (depth 31/32), AA
[ 4704.104491] ata7.00: configured for UDMA/100
[ 4704.120027] ata7: EH complete
[ 4704.120119] scsi 6:0:0:0: Direct-Access     ATA      C300-CTFDDAC256M 0006 PQ: 0 ANSI: 5
[ 4704.120258] sd 6:0:0:0: Attached scsi generic sg3 type 0
[ 4704.120516] sd 6:0:0:0: [sdd] 500118192 512-byte logical blocks: (256 GB/238 GiB)
[ 4704.120554] sd 6:0:0:0: [sdd] Write Protect is off
[ 4704.120556] sd 6:0:0:0: [sdd] Mode Sense: 00 3a 00 00
[ 4704.120568] sd 6:0:0:0: [sdd] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

I rebooted and payed closer attention to the post screen - indeed, one of my precious 256GB C300s deigned to disappear. Despair and panic set in quickly. Many reboots and Google searches later, I was no closer to a solution. In blind rage, I hit the SSD with a hammer.

Suddenly contrite, I realized that I'd possibly voided the SSD's warranty. Worrying that the SSD had been damaged and was liable to short, I plugged it into a spare power supply and turned it on with an ATX power supply tester gadget. Then, I went and did more mundane things for a few hours.

Upon my return, I attached a SATA cable to the SSD and, to much to my relief, found that the drive was detected and seemed to be in perfect working order!

If your C300 fails to be detected after a power failure, don't hit it with a hammer. Just supply it power for a couple hours or perhaps at least 20 minutes. I assume that power loss during certain important operations requires the drive to perform some lengthy internal consistency checks. So, be patient, read a few chapters or watch something on TV, and when you come back, everything will be right as rain. And not destroyed by hammers.

Labels: hardware, SSD