Seagate ST31500341AS 1.5TB

Hi,

just a few days i was happy with my readynas pro running 4.2.1 and using the new seagate 1.5TB drives. No freezing. But now, like all other users i've experienced the same issue. Also disabeling the write cache did help but decreases the performance by about 50% as expected. Let's try to analyse what really happens:

- No freeze during installation (heavy load)
but in normal operation:
- Freeze when in low throughput mode (dvd media viewing, 200-500KB/sec)
- No freeze when load is high (file copy, read and/or write, 25-60MB/sec) )
- No freeze with disabled write cache (load independent)

The question is: Why does a disabled write-cache inhibits freezing during reed-only operation?

In a raid configuration, blocks are spread over several drives (with checksum to the parity drive), so a quite long time of inactivity may result to certain drives in a given raidset.

My best guess:
- Seagate improved 'green power technologie' (see homepage) to let the drive spin down during inactivity, independent of any OS settings.
- Disabeling the write cache keeps more activity to the drives and pretend them from spinning down.

Hints:
- Having a look at the drive temperaures and system fan speed indicates raised power consumption when write cache is disabled.
- freezing period is about 27 sec and much closer to a spin up time than to the required time by a forced flush of the 32MB cache (should take less than 1sec)
- freezing is not (often) reported with vista (because of heavy drive usage by the OS :-) )

So may be it's not a bug but a (very unpleasant) feature.

Dear Jedi-team, any chance to send a code to the drives to prevent them from spinning-down, just for testing?

How often does the freeze occur? Can you make it happen while sitting next to the RAID to listen for a spinup? That would confirm your suspicions.

Freezing delay depends on sytem transfer load. If I start slow transfer out of idle mode (Waiting 15 min, then starting video viewing) first freeze (27sec) occurs after about 10sec (may be there is something cached), followed by a 15sec playback, followed by 27sec freeze... working like a clock for about 5 freezes. Thereafter it gets a little bit mixed but maximum playback is never more than 1-2 minutes followed by a freeze. Assuming the blocks are read round robin out of 5 drives (raidset is 6 drives including parity) there is a maximum time of about 5 minutes for a drive to fall asleep.

Starting the game out of full transfer load (starting media playback during a file copy, then aborting copy) you get the above 'mixed' situation. If I start file copy operation again, freezing disapears.

Unfortunately there is no 'spin up noise', but may I can't listen because of the other spinning drives or it's only a partitial spin down. Have a look at seagate homepage, these drives are supporting various energy saving modes like slow spinning....

May be these drives only believe they have to do wake up and perform the power up sequence, rotating well all time ...?

mraneri wrote:

fearless_fool wrote: ...assuming Seagate comes out with a new firmware rev that resolves the write cache / streaming issue, how does a mere mortal upgrade the firmware in the drives? Seagate's general stance is that firmware is not upgradeable by end users. Of course, an end user can perform the upgrade via a PC, but whether Seagate makes this upgrade available to end users remains to be seen. The standard procedure would be to send the drive back for an RMA replacement, in which case you will get a different refurbished drive. Check out the seagate website for yourself to read their policies.

that may be true, but on multiple occassions they have sent me firmware updates to fix specific issues i contacted them about; for example, when i had 4 drives in a raid array, one with a different fw version. upgrading that drive fixed the issue (see link here viewtopic.php?f=24&t=16600 )

I have this drive (SD 17, UDMA mode 5, NTFS partitions 195+195+1006) as a boot drive attached to Asustek P5K mobo. With 32bit Vista Ultimate SP1 the system randomly but completely freezes up for a couple of minutes (HD LED is lit). After that everything is ok again, no need for rebooting. With XP SP3 there are no problems.

At this moment the drive is to be avoided.

Hope they can fix it with a user programmable FW update (as is the case with the ES drives). Or they are going to get all of them back...

-LaJazz

I have this drive, and had 2 reallocated sector counts pop up since I've had it (~1 month).

It is running SD17 also.

Right now I am only keeping stuff on it as a non-primary backup, but once the issues appear to get worked out I'll eventually fill the nv+ with them and be watching for further problems.

Seagate statement: http://techreport.com/discussions.x/15863

GibberishDriftwood wrote:
Seagate statement: http://techreport.com/discussions.x/15863

I take that statement to indicate that ONLY the affected part numbers: 9JU138-300, 336 with firmware revisions SD15, SD17, or SD18, are experiencing problems. Which leads to the obvious question, when is netgear going to test the 1.5gb drives with SD37 which are reported to be free of these write cache problems??

rstone wrote:

GibberishDriftwood wrote: Seagate statement: http://techreport.com/discussions.x/15863 I take that statement to indicate that ONLY the affected part numbers: 9JU138-300, 336 with firmware revisions SD15, SD17, or SD18, are experiencing problems. Which leads to the obvious question, when is netgear going to test the 1.5gb drives with SD37 which are reported to be free of these write cache problems??

Better question when is seagate gonna fix the problem ;)

GibberishDriftwood wrote:

rstone wrote: GibberishDriftwood wrote: Seagate statement: http://techreport.com/discussions.x/15863 I take that statement to indicate that ONLY the affected part numbers: 9JU138-300, 336 with firmware revisions SD15, SD17, or SD18, are experiencing problems. Which leads to the obvious question, when is netgear going to test the 1.5gb drives with SD37 which are reported to be free of these write cache problems?? Better question when is seagate gonna fix the problem ;)

Well for people like me who didn't jump the gun and purchase the 1.5TB drives, that really isnt a concern, as long as there are 1.5TB drives that do not have a write cache problem. While that sucks for those who did, netgear shouldn't wait for those early adopters to get their firmware before testing the new SD37 drives.