Subject: | |
From: | |
Reply To: | |
Date: | Thu, 25 Apr 2013 18:39:53 -0700 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
On 04/25/2013 02:58 PM, Konstantin Olchanski wrote:
>> joe@george:~$ sudo smartctl -iA /dev/sdd
>> Model Family: SandForce Driven SSDs
>> Device Model: Corsair CSSD-F240GB2
>> Firmware Version: 2.0
>> User Capacity: 240,057,409,536 bytes [240 GB]
>>
>> 171 Program_Fail_Count 0x0032 000 000 000 Old_age Always - 1
>> 174 Unexpect_Power_Loss_Ct 0x0030 000 000 000 Old_age Offline - 45
>> 181 Program_Fail_Count 0x0032 000 000 000 Old_age Always - 1
> Your SSD is toast... Program_Fail_Count non zero is bad. Unexpected power loss count non zero is bad.
>
I agree, it's now two days with Einstein at Home running using 100% CPU
and GPU with no crashes.
The symptoms were certainly strange. To recap:
On average once or twice a day the system would just stop working with
error messages occasionally and complete freeze most of the time. A
power cycle seemed to reset the clock and let me work for another 4 to 8
hours.
Error messages included "too many files open", "i/o error", "bus error"
and probably more.
The SSD was my system disk but swap, /tmp, /usr were on spinning media.
When it worked it did everything I tried, copy files and diff, read all
files, run SMART diagnostic tests.
The SSD was only marginally faster (judging by response time). I don't
think I'm going to replace it with another SSD. However, I have nothing
against SSD, almost 2 years of trouble free operation and a sample size
of one does not lead me to any conclusions about them.
Thanks to the list, I now have more Linux diagnostic tools.
Joe
|
|
|