Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Locked thread
Cron PERLman
Jan 18, 2005

20 4 * * /bin/smokeweed
I smoke the weed at 4:20am
Question regarding the Samsung 850 Pro 250GB drives:

We installed 2 drives to provision an ubuntu server with and when running tests we find the drives work 100% except when we configure both as a raid1 (mdadm software raid) array. According to the system log output, the drive modes start at 6gbps then gradually scale down to 3gbps, finally 1.5gbps before giving a kernel panic.

We're trying to figure out if it is in fact a problem with the drive firmware as we have multiple setups like this with other branded drives, even Samsung 840s.

Has anyone else run into or heard about something like this, or perhaps have access to 850 drives and willing to perform a test to confirm?

Some log output:

code:
[ 3732.178360] ata1.00: exception Emask 0x10 SAct 0x3 SErr 0x400100 action 0x6 frozen 
[ 3732.178385] ata1.00: irq_stat 0x08000000, interface fatal error 
[ 3732.178397] ata1: SError: { UnrecovData Handshk } 
[ 3732.178412] ata1.00: failed command: WRITE FPDMA QUEUED 
[ 3732.178423] ata1.00: cmd 61/08:00:88:c9:7b/00:00:1e:00:00/40 tag 0 ncq 4096 out 
[ 3732.178426] res 40/00:00:88:c9:7b/00:00:1e:00:00/40 Emask 0x10 (ATA bus error) 
[ 3732.178456] ata1.00: status: { DRDY } 
[ 3732.178465] ata1.00: failed command: WRITE FPDMA QUEUED 
[ 3732.178689] ata1.00: cmd 61/08:08:78:ec:3b/00:00:0f:00:00/40 tag 1 ncq 4096 out 
[ 3732.178690] res 40/00:00:88:c9:7b/00:00:1e:00:00/40 Emask 0x10 (ATA bus error) 
[ 3732.179550] ata1.00: status: { DRDY } 
[ 3732.179775] ata1: hard resetting link 
[ 3732.484754] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300) 
[ 3732.486619] ata1.00: supports DRM functions and may not be fully accessable. 
[ 3732.486798] ata1.00: supports DRM functions and may not be fully accessable. 
[ 3732.486822] ata1.00: configured for UDMA/133 
[ 3732.486828] ata1: EH complete
Again, this only happens when configured in software raid.


edit: unrelated, I thought this was a relatively good and informative article with regards to over-provisioning, explaining how it works and why its necessary. Also contains general information on how NAND drives work, definitely worth a read : http://www.lsi.com/downloads/Public/Flash%20Storage%20Processors/LSI_WP_Over_provisioning.pdf

Cron PERLman fucked around with this message at 12:42 on Aug 26, 2014

Adbot
ADBOT LOVES YOU

Cron PERLman
Jan 18, 2005

20 4 * * /bin/smokeweed
I smoke the weed at 4:20am
Weird thing I noticed a few days ago:

Windows 8.1, 8GB RAM.
25 day old Samsung 840 Evo 250GB as system drive with RAPID enabled.
Samsung magician reports 4.88TB written.

That's ~195GB per day, which is impossible as I only use the PC for 3-4 hours a day max, with no heavy writes going on. Most of the large file writes were performed weeks ago when I installed Windows and steam + a few games with only 70GB used space on the drive.

I opened up Resource Monitor and noticed that after a few minutes of inactivity my system just bursts data to c:\pagefile.sys for seemingly no reason as I always have about 40% ram free when the PC is idling, RAPID taking up the majority of used ram.

Since disabling RAPID I haven't seen it writing to the pagefile, yet. Anyone else experienced this? Surely the behaviour isn't normal and I'd certainly have had a dead SSD in a year's time if I hadn't picked up on it.

Only registered members can see post attachments!

Cron PERLman fucked around with this message at 21:35 on Sep 22, 2014

Cron PERLman
Jan 18, 2005

20 4 * * /bin/smokeweed
I smoke the weed at 4:20am

HalloKitty posted:

Isn't that really quite a lot of writes, though? My Samsung 830 256GB has been on for 5336 hours, and only reports 3.58TiB written.

Yeah I'm pretty sure that a normal user shouldn't do more than 10-20GB, which is why my 200GB per day is bugging me.

  • Locked thread