What is the next SSD victim ?
Printable View
What is the next SSD victim ?
Kingston SSDNow 40GB (X25-V)
353.66TB Host writes
Reallocated sectors : 8
MD5 OK
--
This time it lasted for ~33 hours.
It disconnected again last night 07:01 CET, so about 6 hours ago and again while deleting files, this time all files were deleted.
I'll move it to another rig this weekend just to see what happens. (another Z68)
Corsair Force 3 120GB
01 120/50 (Raw read error rate)
05 2 (Retired Block count)
B1 40 (Wear range delta)
E6 100 (Life curve status)
E7 83 (SSD Life left)
E9 73936 (Raw writes)
F1 98538 (Host writes)
power on hours : 281.
I've lost about 18 hours in the past week to disconnects... It disconnected again again last night about an hour after I went to bed. The worst time that it can happen.
This weekend I'm going to swap out motherboards for an H67 and just hope that helps.
I might try one of the other (non SB/PCH) ports on the current rig, not sure what it is but it's one of the typical add-on controllers.
If that doesn't work out I'll try the lsi 9211, if SMART still works that is.
It's annoying alright but I'm not giving up on the drive.
No, I won't be giving up either, but that doesn't mean I'm at all pleased with some of the nonsensical explanations and voodoo magic solutions. I think this is how many of the Vertex 3 owners felt after the products' launch, and part of me just wonders if the some of the SF 2281 controllers are just faulty. If worse comes to worse, I'll swap it out under warranty and try with another... But that's the worst case scenario.
Until I switch to another motherboard, I've switched to a SATA II port, so we'll see if it helps. I tried it last week, but just to see what kind of speeds I'd achieve.
With the P67's 6gbps port I was getting 122 - 124 MBs avg.
With the P67's 3gbps port I'm getting 117 - 119 MBs avg.
With a C2D ICH8m SATA II laptop I was getting about 94 - 96 MBs.
I can live with 118MBs if it means I can make it through the night, but I'm not sure that Sata II ports are any less prone.
I would be interested to know if the issues are more hardware related or there's a combination of hardware-software (drivers with issues that occur under certain conditions). For ruling out software, you could install a Linux distribution and run any program that would generate continuous writes, as SMART parameters are already tracking everything. I know this might be a too far suggestion but it might worth trying.
Update
m4:
561.2012 TiB
2073 hours
Avg speed 88.88 MiB/s.
AD gone from 34 to 30.
P/E 9818.
MD5 OK.
Still no reallocated sectors
Attachment 120636Attachment 120637
Kingston V+100
80.6501 TiB
323 hours
Avg speed 79.01 MiB/s.
AD gone from 129 to 116.
P/E ?.
MD5 OK.
Attachment 120638Attachment 120639
While I was at work ASU reported errors and froze. I could open folders and files on the SSD but ASU had no response. I shut it down and tried to open it again. Now the startscreen came up but the program froze while trying to load information. I powered down the rig and tried a restart. Next problem now was that the JMicron optionrom couldn't find the Kingston so I disconected it and took a new restart.
I moved the Kingston til SB850 and got help for Anvil to restore the log and now all runs normal again. The MD5 check showed no errors so I don't know what caused all of this.
Here is a screenshot of ASU with the error message.
Attachment 120640
Kingston SSDNow 40GB (X25-V)
355.18TB Host writes
Reallocated sectors : 8
MD5 OK
36.97MiB/s on avg (~12 hours)
--
Corsair Force 3 120GB
01 90/50 (Raw read error rate)
05 2 (Retired Block count)
B1 42 (Wear range delta)
E6 100 (Life curve status)
E7 82 (SSD Life left)
E9 77332 (Raw writes)
F1 103064 (Host writes)
107.48MiB/s on avg (~12 hours)
power on hours : 293
@sergiu
Not very likely, it would be the last thing I'd try, if there is something to the SF/SB platform mix that stinks I'll make sure to test on other SB computers and then go down the X58 route as well before trying another OS/platform.
Mushkin Chronos Deluxe 60 Update
05 0
Retired Block Count
B1 19 (Up from 18)
Wear Range Delta
F1 74681
Host Writes
E9 57563
NAND Writes
E6 100
Life Curve
E7 72
Life Left
Average 116.41MBs. P67 SATA II
119 Hours Work Time
12GiB Minimum Free Space
SSDlife expects 18 days to 0 MWI
Attachment 120644
I've lost a lot of time with the issues I've had. I'm running it on SATA II until it crashes again, then I'm switching motherboards to see if it helps.
EDIT
I've heard that the Marvel controlled SATA III ports have problems too, but instead of switching to another 1155 board, I think I'll get a PCIe SATA III controller. If that doesn't work, I'll buy some kind of HBA. If that doesn't work, I'm flying to Sandforce HQ to yell at someone.
Looks like the SATA II ports don't help either; I just got my first disconnect on the P67's SATA II ports.
I've increased PCH voltage... not sure if that'll help...
Anvil, what kind of Z68 board are you using?
96.10 hours
228.8066 TiB written
57.86 MB/s
MD5 ok
05: 0
B1: 130
E7: 10%
E9: 186624
EA/F1: 235456
F2: 384
@christopher
It's currently running of an ASRock Extreme4 Z68, either I will change ports or it will run off an Asus M4E-Z for the weekend only.
@synbiosvyse
96 hours can't be right, that is just 4 days and you've been running for months? (or is it this session only)
Otherwise it looks great, no LTT and no reallocated sectors, one should think there would have been movements on 05 by now.
Looks like the M4 still has some way to go.
Kingston SSDNow 40GB (X25-V)
356.41TB Host writes
Reallocated sectors : 9 (from 8)
MD5 OK
35.08MiB/s on avg (~23 hours)
--
Corsair Force 3 120GB
01 95/50 (Raw read error rate)
05 2 (Retired Block count)
B1 44 (Wear range delta)
E6 100 (Life curve status)
E7 81 (SSD Life left)
E9 80407 (Raw writes)
F1 107163 (Host writes)
107.51MiB/s on avg (~23 hours)
power on hours : 304
Update
m4:
565.0567 TiB
2085 hours
Avg speed 89.57 MiB/s.
AD gone from 30 to 27.
P/E 9880.
MD5 OK.
Still no reallocated sectors
Attachment 120652Attachment 120653
Kingston V+100
83.8624 TiB
335 hours
Avg speed 77.14 MiB/s.
AD gone from 116 to 114.
P/E ?.
MD5 OK.
Attachment 120654Attachment 120655
There have been no problems since yesterday so I'm crossing my fingers that it will continue.
Yeah that's since I switched back to 46%. I had been running 0-Fill for a while at the beginning, so I ran 100% to help offset that. Personally I find it a little confusing that the app resets the timer even if you set it to count totals. So if you stop the app, it still continues to carry over the data written, but then it resets the timer. Have you considered changing this? Or perhaps at least have two timers, a global and session only.
Overall I've been very impressed with this drive. The speeds are still almost the same as day one, and no reallocated sectors. The MWI has been stable at 10% for a little while now. I will continue to monitor this and note when it drops.
WOW, imagine what the results of the C300 will be if the M4 got this far ! Really impressive. and no SF-like bugs of random dissapearances etc.
Dang she's going to crack 10k on that M4. 650TB no issues I'd say at least before it runs out on SSDLife Pro.
At least hes not testing a 128Gig drive... that thing would be going for months more!
That assumes the 34nm NAND is better, and maybe it is, but from what I've seen so far it might be a moot point -- anything's better than random dropouts.
I'm rapidly starting to lose all the goodwill I feel towards the Mushkin. Not that it's Mushkin's fault -- It's probably Sandforce's, but that doesn't make it any less annoying. It dropped out last night again. I'm not really sure that anything I try is going to help.
Mine is past 30 hours, I'm sort of expecting it to disconnect within a few hours. (that would fit most of the other disconnects)
I have seen this problem with the SF-1200's last year (on Corsair's). Disabling TRIM did not help in most cases. Have you tried disabling TRIM? I just don't think it will fix the problem. I think there's a worse underlying problem with the firmware.
I don't think it will either. I tried it for a little while, but from what I've seen with users with the exact same problem, it won't help. What I was saying was, if it did help but dropped avg speed 25%, it would still be a net gain. Every night, at some point after I go to sleep, it drops out. So I'm losing several hours every night. My backup AM3 rig only has SATA II, but I'm considering switching to it as well instead of another 1155 board.
I only started looking into TRIM because the first three crashes happened during the delete section of the loop. Since then it's happened seemingly at random. I'm not going to give up, but I'm thinking RMA. I'm also thinking about trying a Samsung 470 with double the spare area to see if it would last twice as long, but not until I get the SF sorted out.
It's not worth it to disable TRIM in my opinion.
You may have mentioned this before but what SATA controller are you using again? If you have an Intel SATA II (like an ICH7-ICH10 or something) then see if that works.
I would definitely return it and get a refund and get a whole new drive, or at the very least RMA it and get a new drive of the same one to see if there is any difference.
Yeah, I tried it on my P67's SATA II port. While I didn't lose much speed, it didn't help either, and disconnected just the same. Right now none of these drives are in stock and so I wouldn't think of RMAing until I can get another one. Plus, if I RMA it, and the drive gets tested it will work, so then they're going to send me the drive back. I'm not sure how best to work this out, but there is definitely something wrong.
Next time it crashes, I'm putting it in the AM3 SATA II ports on my backup system.
EDIT
If anyone has any ideas of another drive to test, I'm all ears.