Jump to content
RESET Forums (homeservershow.com)
Alex.Covecube

Successfully Enabled Port Multiplier on eSATA

Recommended Posts

Puulima

Update: My Richcopy batch completed and coincidentally so did the AMD_SATA errors?? The Event Viewer shows that error reurring every few minutes right up until the RichCopy job ended. Then no more errors? Right now I'm running concurrent WRITE tests to all of my drives and no errors and good throughput...and no errors in the Event Viewer log.

 

Should I post this somewhere else? Either in another forum on this site or elsewhere? I don't know who to seek assistance from...

Share this post


Link to post
Share on other sites
ikon

Would really like to see a comparison copy done with RoboCopy to, you know, see if the errors reoccur.

Share this post


Link to post
Share on other sites
Puulima

Would really like to see a comparison copy done with RoboCopy to, you know, see if the errors reoccur.

 

Well here's what I did - and this doesn't make sense to me.

 

I copied 80GB of data over from the Storage Spaces to the same target drive I've been using - using Windows default copying method - no issues, Average Speed was in the mid 20's - and no errors (or other events) in the event Viewer window. I then deleted the copied data on the target and repeated with RichCopy - approx 3 minutes into the job, RichCopy freezes and stays that way for almost 3 minutes...as soon as it starts back up copying - the same "amd_sata" error shows up in the Event Viewer. A minute and 40 seconds later, the copying stalls again...more error messages in Even viewer. As I type it's frozen for the third time and wait for it....yup, as soon as it starts back up Event Viewer indicates "new events avaialable" - instantaneously with the re-start of copying!

 

So I kill that copy job, delete the copied data and change the RichCopy settings so it's only copying one thread at a time (just a hunch) - and guess what? - No issues whatsoever copying all 80 GB - no errors or events.

 

I repeated these test scenarios and same results - so while not an extensive testing sample - a little hard to believe its a coincidence.

 

I don't have the technical skills or knowledge to get under the hood here...but it seems that I can use RichCopy as long as I stick with a single thread - at least for my immediate drive backup needs - or do you think I'm better off just using xcopy? I've given up the plan to consolidate the data (and therefore eliminate the dupes) and I'm just making drive dupes to be safe. I found a "how to" copy your WHS v1 data 9icnluding duped files) over to the file structure for DrivePool here:

 

http://forum.covecube.com/discussion/574/how-to-migrating-from-whs-v1039s-drive-extender

 

So That will be my plan when I get the WHS 2011 OS setup (software shipped out this afternoon).

 

I wonder if there's somewhere I can report this issue with RichCopy - but I think I'll test it on my Windows 7 machine also first.

Share this post


Link to post
Share on other sites
Puulima

I don't think this is (solely) a RichCopy issue - I'm now copying from one drive to another in the N40L itself and that's working fine 0 even with 3 threads. no amd_sata errors. I'm going to lose the Storage Spaces tomorrow and test this again with the individual drives in the external enclosure and see what happens.

Share this post


Link to post
Share on other sites
ikon

My experience from yesterday matches up pretty well with yours. I attached a 1 TB Red drive via eSATA and used RichCopy to copy about 650GB of data to a 2x2TB Red drive internal mirrored array. I had previously copied the same data to the 1 TB Red drive using RoboCopy.

 

I used 10 directory threads for RichCopy. I had to leave the site before the copy job was complete, but my impression was that it wasn't progressing any faster and, indeed, seemed to be going slower than the RoboCopy job. This is particularly disturbing because the RoboCopy job was done on my old Acer H340. I would expect the N40L, if anything, to be faster hardware than the H340. So, if anything, I would have expected RichCopy on new hardware to have proceeded significantly faster than RoboCopy on older hardware.

 

I will know more tomorrow when I get a chance to look at the (hopefully) completed RichCopy job.

Share this post


Link to post
Share on other sites
Puulima

Steering this thread back onto the specific topic: If I'm installing WHS 2011 on my N40L - sounds like I use the drivers in the \W764A folder based on Alex's comment in the first posting:

 

I extracted the EXE using 7-ZIP and found the correct driver in \12-10_vista_win7_win8_32-64_sb\$_OUTDIR\Packages\Drivers\SBDrv\SB7xx\AHCI\W764A. Just open device manager and update the driver manually pointing it to this location. This was for Windows Server 2008 R2, so all WHS variants will use the same driver.

 

Nothing else should change though - IE: I've already loaded on the modded BIOS and tweaked all the reccomended settings (and PM is working in Windows 8 right now).

Share this post


Link to post
Share on other sites
UhClem

My experience from yesterday matches up pretty well with yours. I attached a 1 TB Red drive via eSATA and used RichCopy to copy about 650GB of data to a 2x2TB Red drive internal mirrored array. I had previously copied the same data to the 1 TB Red drive using RoboCopy.

 

I used 10 directory threads for RichCopy. I had to leave the site before the copy job was complete, but my impression was that it wasn't progressing any faster and, indeed, seemed to be going slower than the RoboCopy job. This is particularly disturbing because the RoboCopy job was done on my old Acer H340. I would expect the N40L, if anything, to be faster hardware than the H340. So, if anything, I would have expected RichCopy on new hardware to have proceeded significantly faster than RoboCopy on older hardware.

 

I will know more tomorrow when I get a chance to look at the (hopefully) completed RichCopy job.

It sounds like you're experiencing the effects of (disk) "head thrashing". If multiple copy threads are operating on the same drive, these threads are competing for that drive's disk head assembly, taking them "off track" for the other thread. Keeping multiple drives active for a large copy operation is good, but ONLY if the "thread master" strives to avoid "head contention". Given the origin of RichCopy (MSFT), I'm inclined to believe it is implemented correctly. Maybe there is a different explanation for your situation.

 

[To: Puulima] However, even with a non-head-thrashing RichCopy, using multiple drives, concurrently, within a CBS-controlled port multiplier enclosure, can be counter-productive. See my above report results for details/explanation. That does not excuse, but might be the basis for, the driver warnings you got.

Share this post


Link to post
Share on other sites
ikon

It sounds like you're experiencing the effects of (disk) "head thrashing". If multiple copy threads are operating on the same drive, these threads are competing for that drive's disk head assembly, taking them "off track" for the other thread. Keeping multiple drives active for a large copy operation is good, but ONLY if the "thread master" strives to avoid "head contention". Given the origin of RichCopy (MSFT), I'm inclined to believe it is implemented correctly. Maybe there is a different explanation for your situation.

 

[To: Puulima] However, even with a non-head-thrashing RichCopy, using multiple drives, concurrently, within a CBS-controlled port multiplier enclosure, can be counter-productive. See my above report results for details/explanation. That does not excuse, but might be the basis for, the driver warnings you got.

 

Head Thrashing was one of my considerations too. Those poor heads, jumping back and forth to meet the demands of writing so many different files; it's just not fair :) I suspect part of the issue is that the Red, Greens, and other lower RPM, low power drives can't really handle such high throughput demands. They are, after all, a fair bit above the ordinary.

Share this post


Link to post
Share on other sites
Puulima

Steering this thread back onto the specific topic: If I'm installing WHS 2011 on my N40L - sounds like I use the drivers in the \W764A folder based on Alex's comment in the first posting: He noted "all WHS variants will use the same driver"

 

Just confirming that I'm understanding this correctly.

 

Also - old WHS v1 never recognized my External 4 Drive enclosure (Sans Digital 4 Drive) - so doing a server Re-installation left me rebuilding all the shares by re-copying all the data. When I boot up to install WHS 2011 on my N40L (with the BIOS mods) the window where you select the drive to install on "sees" the first drive in the enclosure - so better off than before already - but since the OS isn't running yet the other drives are NOT SEEN.

 

Does anybody know what will happen in the case of a re-installation as far as data on those other 3 drives? will I be in the same bad place as before? My thinking it that if WHS can't see the other drives in the external enclosure until the DRIVERS are loaded - which won't be until the OS is up and running and I can install them - then it won't be able to rebuild the data Pool. Probably not an issue for the initial install but I'm thinking perhaps I should try an load those drivers via USB (clicking the Load Drivers button) and test now (probably answering my own question here - but wanted to bounce my thoughts off of y'all).

 

Here's another wrinkle - if I'm ultimately using DrivePool and all my data is stored under that add-in - how does WHS 2011 handle all the data on a Server Re-installation? I went googling but no specific threads found yet.

 

As always, thanks to all.

 

(Mods - perhaps this should be moved to another forum topic? But it is related to the PM functionality so I was torn).

Share this post


Link to post
Share on other sites
ikon

I generally install the OS with only 1 drive in the system. This avoids a number of potential issues.

 

I don't know for certain, but I would expect, when you load the drivers for the enclosure after WHS2011 is running, it will suddenly see the other drives, and also see any data already on those drives. I would not expect it to corrupt or delete any data, but I would always have at least 1 complete backup of all the data, you know, just in case... ;)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...