Are users of SATA port multipliers affected by this as well?

 

Probably not, but that is a good question.

 

I have a port multiplier and this issue is absolutely killing me.

 

Silicon Image 3726 with a Sil 3132-based card.

 

Please provide more details of your config and select/copy/paste the relevant lines of the Devices page - I'm curious if the set of drives on one controller is > 2 (probably is for command-based port multipliers).

 

parity device:  pci-0000:00:1f.2-scsi-0:0:0:0 (sdb) ata-SAMSUNG_HD154UI_S1Y6J1KS738525

disk1 device: pci-0000:00:1f.2-scsi-1:0:0:0 (sdc) ata-ST31000340AS_3QJ019MW

disk2 device: pci-0000:00:1f.2-scsi-4:0:0:0 (sdd) ata-ST31000340AS_3QJ022CD

disk3 device: pci-0000:00:1f.2-scsi-5:0:0:0 (sde) ata-ST31000340AS_3QJ03CS7

disk4 device: pci-0000:05:00.0-scsi-0:4:0:0 (sdl) ata-ST31000340AS_3QJ01EGG

disk5 device: pci-0000:05:00.0-scsi-0:3:0:0 (sdk) ata-SAMSUNG_HD154UI_S1Y6J1KS738524

disk6 device: pci-0000:05:00.0-scsi-0:1:0:0 (sdi) ata-ST3500630AS_5QG1QL3A

disk7 device: pci-0000:05:00.0-scsi-0:2:0:0 (sdj) ata-WDC_WD5000AAKS-00YGA0_WD-WCAS81016877

disk8 device: pci-0000:05:00.0-scsi-0:0:0:0 (sdh) ata-ST3500630AS_6QG39S2G

disk9 device: pci-0000:03:00.0-scsi-1:0:0:0 (sdg) ata-WDC_WD5000YS-01MPB0_WD-WMANU1411028

 

Parity and disks 1-3 are on the onboard ICH9-based SATA ports.  Disks 4-8 are on the 3726/3132 combo.  Disk9 and a currently unused Disk10 are both on an additional 2-port 3132-based PCIe controller.

 

Motherboard is an Intel DG965RYCK (i965 chipset/ICH9).  Celeron 430 (ran a Celeron E1200 dual core for quite a while with same results) and 4GB DDR2-800.  I'm pretty certain I switched to AHCI mode at some point, but I will double check and note here if it is different.

I believe Reiser 4 was supposed to remove BKL

 

In a mailing list post from July 2009, Edward Shishkin wrote that in the coming autumn, they would start exploring the opportunity of getting Reiser4 into the main Linux kernel.[14] In a November 2009 interview[15] to Phoronix he said he is going to publish plug-in design document for independent expert review. He is currently aiming for USENIX Annual 2010. If all goes well, Reiser4 may enter mainline Linux kernel around version 2.6.36.

As you can see from this this thread

Reiserfs 3.6 is currently being patched against the kernel 2.6.31 tree to eliminate the BKL.  It will probably not be necessary to wait for ReiserFS 4.0 before a version is included in the kernel that has it removed.  (If I'm reading it right, it might be as early as 2.6.32 or 2.6.33)

 

Joe L.

Hi!

 

Some change between 4.5.8 and 4.5.11 affected negatively my reading speed from the unraid server:

 

In 4.5.8, I copy (with Teracopy, set to read back the whole thing afterwards and verify the crc) a 7GB movie folder to the server, and it writes it (to a disk outside the protected aray) with ~30MB/s and then reads it back and verifies the CRC with ~50MB/s.

 

In 4.5.11, I copy exactly the same folder to the same disk, and it writes it with ~30MB/s and then reads it back and verifies the CRC with ~20MB/s.

 

I restored the 4.5.8 bzimage and bzroot to the flash key, rebooted the server, and the same operation is back to ~30MB/s write, then ~50MB/s read back and verify.

 

What could possibly be causing this decrease in read speeds from the server?

Could this have something to do with changes in Samba?

 

 

What could possibly be causing this decrease in read speeds from the server?

Could this have something to do with changes in Samba?

 

Please download the attached file and place in 'config' directory of Flash, then Stop/Start array & let me know if this makes any difference.

What could possibly be causing this decrease in read speeds from the server?

Could this have something to do with changes in Samba?

 

Please download the attached file and place in 'config' directory of Flash, then Stop/Start array & let me know if this makes any difference.

 

YES!  That did it! 

 

Now 4.5.11 is back to the same speeds as 4.5.8.

 

BTW, before the patch, I was watching the copy with htop in telnet, and I was wondering why there are two smbd processes running together and eating lots of cpu together for the same copying operation...

 

Good catch. Thanks.

 

[edit]:

If you tell me that this "asynchronous I/O" stuff is supposedly bringing better stability and responsiveness, then I may reconsider, and learn to live with the slower read speeds.

 

 

That means that Samba Async I/O isn't a benefit for everyone. I don't suppose there's an easy way of making this selectable via emhttp?

 

 

That means that Samba Async I/O isn't a benefit for everyone. I don't suppose there's an easy way of making this selectable via emhttp?

 

Async I/O is probably beneficial for everybody else but me. My build is seriously underpowered.

 

As of making it selectable, Tom just showed us how we can disable it if we need.

 

 

....which got me thinking...

I've accumulated a whole bunch of tweaks which are highly beneficial only for weak, ultra-low-voltage processors.

So may be a good idea if I put them together in a page for the wiki.

Agreed, but for those who are not as tech savvy they may appreciate it being controlled that way.

 

Or better yet perhaps in 5.x / 6.x there would be different system profiles to select from with the various tweaks/settings combined.

....which got me thinking...

I've accumulated a whole bunch of tweaks which are highly beneficial only for weak, ultra-low-voltage processors.

So may be a good idea if I put them together in a page for the wiki.

 

If I can't straighten out my upgraded board (that crashes when starting the array), I'll probably be very interested in those tweaks.  Think it can get down to a 1100mhz Athlon XP?

My config, which does exhibit the media stall:

 

pci-0000:00:0e.0-scsi-1:0:0:0 (sdn) ata-SAMSUNG_HD154UI_S1Y6J1KS744219

disk1 device: pci-0000:00:0e.0-scsi-2:0:0:0 (sdo) ata-WDC_WD7500AAKS-00RBA0_WD-WCAPT0618609

disk2 device: pci-0000:00:0e.0-scsi-3:0:0:0 (sdp) ata-ST3750640AS_3QD078ZS

disk3 device: pci-0000:04:00.0-scsi-0:0:0:0 (sdj) ata-WDC_WD10EADS-00L5B1_WD-WCAU49708236

disk4 device: pci-0000:04:00.0-scsi-0:1:0:0 (sdk) ata-WDC_WD10EADS-00L5B1_WD-WCAU49710186

disk5 device: pci-0000:04:00.0-scsi-0:2:0:0 (sdl) ata-WDC_WD10EADS-00L5B1_WD-WCAU49785711

disk6 device: pci-0000:01:08.0-scsi-2:0:0:0 (sdc) ata-WDC_WD10EACS-00ZJB0_WD-WCASJ0974421

disk7 device: unassigned

disk8 device: pci-0000:01:09.0-scsi-0:0:0:0 (sde) ata-WDC_WD10EACS-00ZJB0_WD-WCASJ0969668

disk9 device: pci-0000:01:09.0-scsi-1:0:0:0 (sdg) ata-WDC_WD10EACS-00ZJB0_WD-WCASJ1565983

disk10 device: pci-0000:01:08.0-scsi-1:0:0:0 (sdb) ata-WDC_WD10EACS-00ZJB0_WD-WCASJ1539830

disk11 device: pci-0000:01:08.0-scsi-0:0:0:0 (sda) ata-WDC_WD10EAVS-00D7B1_WD-WCAU48311113

disk12 device: pci-0000:01:09.0-scsi-2:0:0:0 (sdh) ata-WDC_WD10EAVS-00D7B1_WD-WCAU48302382

disk13 device: pci-0000:01:09.0-scsi-3:0:0:0 (sdi) ata-WDC_WD10EADS-00L5B1_WD-WCAU49767827

disk14 device: pci-0000:01:08.0-scsi-3:0:0:0 (sdd) ata-SAMSUNG_HD103UJ_S13PJ9BQC03609

disk15 device: pci-0000:00:0e.0-scsi-0:0:0:0 (sdm) ata-WDC_WD15EADS-00H7B0_WD-WCAUP0004797

 

The drives are in transition and will be reassigned soon - to be more consistent between slot number and controller/physical location.

I also have the media stall during directory listings:

 

pci-0000:00:1f.2-scsi-0:0:0:0 (sda) ata-WDC_WD15EADS-00H7B0_WD-WCAUP0024759

disk1 device: pci-0000:00:1f.2-scsi-1:0:0:0 (sdb) ata-WDC_WD10EADS-00L5B1_WD-WCAU45051875

disk2 device: unassigned

disk3 device: pci-0000:00:1f.2-scsi-3:0:0:0 (sdc) ata-WDC_WD7500AAKS-00RBA0_WD-WCAPT0751279

disk4 device: pci-0000:00:1f.2-scsi-4:0:0:0 (sdd) ata-WDC_WD5000AAKS-00YGA0_WD-WCAS80541016

disk5 device: pci-0000:00:1f.2-scsi-5:0:0:0 (sde) ata-WDC_WD10EADS-65L5B1_WD-WCAU48062314

disk6 device: pci-0000:04:04.0-scsi-0:0:0:0 (sdg) ata-WDC_WD7500AACS-65D6B0_WD-WCAU80040406

disk7 device: pci-0000:04:04.0-scsi-3:0:0:0 (sdj) ata-WDC_WD10EADS-65L5B1_WD-WCAU48049990

disk8 device: pci-0000:04:04.0-scsi-2:0:0:0 (sdi) ata-WDC_WD7500AACS-00D6B0_WD-WCAU50028552

cache device: pci-0000:04:04.0-scsi-1:0:0:0 (sdh) ata-Maxtor_7L300S0_L603LVAG

 

Think it can get down to a 1100mhz Athlon XP?

Yep.  Mine is at 1000MHz.  And it doesn't even have any L2 cache.

The box is drawing ~30 Watts from the wall with 6 disks spun down.

I also have the spin up problem

 

parity device: pci-0000:00:1f.2-scsi-0:0:0:0 (sda) ata-SAMSUNG_HD103UJ_S13PJ1EQ504937

disk1 device: pci-0000:00:1f.2-scsi-1:0:0:0 (sdb) ata-SAMSUNG_HD103UJ_S13PJ1EQ504936

disk2 device: pci-0000:00:1f.2-scsi-2:0:0:0 (sdc) ata-SAMSUNG_HD103UJ_S13PJ1BQA07919

disk3 device: pci-0000:00:1f.2-scsi-4:0:0:0 (sde) ata-SAMSUNG_HD103UJ_S13PJ1EQ504935

disk4 device: pci-0000:00:1f.2-scsi-5:0:0:0 (sdf) ata-SAMSUNG_HD103UJ_S13PJ90Q909841

disk5 device: pci-0000:00:1f.2-scsi-3:0:0:0 (sdd) ata-WDC_WD10EADS-00M2B0_WD-WCAV51184422

 

I also have the spin up problem

 

parity device: pci-0000:00:1f.2-scsi-0:0:0:0 (sda) ata-SAMSUNG_HD103UJ_S13PJ1EQ504937

disk1 device: pci-0000:00:1f.2-scsi-1:0:0:0 (sdb) ata-SAMSUNG_HD103UJ_S13PJ1EQ504936

disk2 device: pci-0000:00:1f.2-scsi-2:0:0:0 (sdc) ata-SAMSUNG_HD103UJ_S13PJ1BQA07919

disk3 device: pci-0000:00:1f.2-scsi-4:0:0:0 (sde) ata-SAMSUNG_HD103UJ_S13PJ1EQ504935

disk4 device: pci-0000:00:1f.2-scsi-5:0:0:0 (sdf) ata-SAMSUNG_HD103UJ_S13PJ90Q909841

disk5 device: pci-0000:00:1f.2-scsi-3:0:0:0 (sdd) ata-WDC_WD10EADS-00M2B0_WD-WCAV51184422

 

 

From what Tom said earlier you should not be having this problem, if I understood him correctly.

 

Please provide a little more details for us and Tom.  Your Motherboard, amount of RAM, version of unRAID (I assume you are running the current beta since this is the beta thread), and anything else you can think of.

It does not look like a shared IRQ issue.

As far as IRQ's being shuffled, anything above 15 (I believe) is programmed in software

Although I don't know what PCI-MSI-edge is.

 

PCI-MSI is an attempt in PCI 2.3 and 3.0 devices to remove the requirement for hardware interupts altogether. MSI = Message Signalled Interrupt.  In order to reduce pin count the document claims (however all CPU manufacturers seem to be increasing pin counts not reducing them (Intel are upto socket 1366 vs amd socket 940). Chuckle I did. PCI-e uses INTx emulation and all INTx capable devices should be able to share interupts, therfore all PCI-e devices should be capable of sharing Interupts. MSI is a message based interupt service using a memory write to the APIC. MSI-X (extensions) allows 32 vectors per device to be requested. This looks to be a mechanism to allow MSI based IRQs to be serviced/directed to specific CPUs. Although I'm not 100% sure I'm reading that right...

 

-edge means the apic is using edge triggering rather than level. Basic MSI how-to is here:

 

http://lwn.net/Articles/44139/

or

http://devresources.linux-foundation.org/dev/robustmutexes/src/fusyn.hg/Documentation/MSI-HOWTO.txt

 

Grabbing a cat /proc/interrupts when the system doesn't autneg 1000Mb/s would be very interesting, also if the link drops to 100Mb/s. It would indicate an issue with MSI implementation in the driver or a lack of available MSI resources. The device should then switch to INTx behaviour, it will be interesting to see what INT it drops too.  

 

 

Thanks. I will definitely check the interrupts again the next time I see a problem. So far it has been reliably reconnecting at 1000 with every WOL. This is consistent with past behavior. i.e., When it works at start-up, it continues to work.

 

Interrupts are assigned exactly the same way in "failure mode", i.e., when a reboot yields only a 100Mb/s connection. So I'm beginning to feel less and less that it is a linux issue and more and more that the hardware is simply flaky.

 

I pulled the network cable and reconnected and it negotiated 1000Mb/s. (14:41:17 in attached syslog). Went into S3 sleep, then WOL and reconnected A-OK at 1000 (@15:31:41). Once it's working it seems stable. Problem occurs only on startup - I think!

 

Maybe all the different behaviors I noted between beta 6, 8, 10-11 were just random and had nothing to do with the linux core or driver changes.

 

Seemed to me I repeat tested enough to be sure, but now I'm having second thoughts.

 

 

OK back to controller situation.

Is there anyone with the Supermicro 8 Port controller who can post the devices and provide results. Might be good for gathering information.

 

This is my configuration (#2 AOC-SAT2-MV8). I have never experienced video stuttering when accessing other disks.

 

parity device:  pci-0000:00:1f.2-scsi-1:0:0:0 (sdq) ata-ST31500341AS_9VS2L8M9

disk1 device: pci-0000:03:02.0-scsi-0:0:0:0 (sdg) ata-WDC_WD10EACS-65D6B0_WD-WCAU41723685

disk2 device: pci-0000:03:02.0-scsi-1:0:0:0 (sdh) ata-WDC_WD10EACS-65D6B0_WD-WCAU41723734

disk3 device: pci-0000:03:02.0-scsi-2:0:0:0 (sdj) ata-WDC_WD10EACS-65D6B0_WD-WCAU41775326

disk4 device: pci-0000:03:02.0-scsi-3:0:0:0 (sdk) ata-WDC_WD10EACS-65D6B0_WD-WCAU41724444

disk5 device: pci-0000:03:02.0-scsi-4:0:0:0 (sdl) ata-WDC_WD15EADS-00P8B0_WD-WCAVU0229967

disk6 device: pci-0000:03:02.0-scsi-5:0:0:0 (sdm) ata-WDC_WD15EADS-00S2B0_WD-WCAVY0840969

disk7 device: pci-0000:03:02.0-scsi-6:0:0:0 (sdn) ata-WDC_WD15EADS-00P8B0_WD-WMAVU0509272

disk8 device: pci-0000:03:02.0-scsi-7:0:0:0 (sdo) ata-WDC_WD15EADS-00P8B0_WD-WMAVU0510079

disk9 device: pci-0000:03:01.0-scsi-0:0:0:0 (sda) ata-WDC_WD15EADS-00P8B0_WD-WMAVU0509105

disk10 device: pci-0000:03:01.0-scsi-1:0:0:0 (sdb) ata-WDC_WD15EADS-00P8B0_WD-WMAVU0512762

disk11 device: pci-0000:03:01.0-scsi-2:0:0:0 (sdc) ata-WDC_WD6401AALS-00L3B2_WD-WMASY3289016

disk12 device: pci-0000:03:01.0-scsi-3:0:0:0 (sdd) ata-WDC_WD6400AAKS-22A7B0_WD-WCASY0757715

disk13 device: pci-0000:03:01.0-scsi-4:0:0:0 (sde) ata-WDC_WD6400AAKS-22A7B0_WD-WCASY0708589

disk14 device: pci-0000:03:01.0-scsi-5:0:0:0 (sdf) ata-WDC_WD6400AAKS-22A7B0_WD-WCASY0681724

[...]

cache device:  pci-0000:00:1f.2-scsi-0:0:0:0 (sdp) ata-WDC_WD6400AAKS-22A7B0_WD-WCASY0684515

I also experience media stall issue

 

parity device: pci-0000:00:1f.2-scsi-4:0:0:0 (sdg) ata-WDC_WD1001FALS-00J7B0_WD-x

disk1 device: pci-0000:00:1f.2-scsi-0:0:0:0 (sdc) ata-WDC_WD1001FALS-00J7B0_WD-x

disk2 device: pci-0000:00:1f.2-scsi-2:0:0:0 (sde) ata-WDC_WD1001FALS-00K1B0_WD-x

disk3 device: pci-0000:00:1f.2-scsi-3:0:0:0 (sdf) ata-ST31000528AS_x

disk4 device: pci-0000:00:1f.2-scsi-1:0:0:0 (sdd) ata-WDC_WD1001FALS-x

disk5 device: pci-0000:00:1f.2-scsi-5:0:0:0 (sdh) ata-SAMSUNG_HD103UJ_x

disk6 device: pci-0000:02:00.0-scsi-1:0:0:0 (sda) ata-SAMSUNG_HD103UJ_x

cache device: pci-0000:01:00.0-ide-0:0 (hda) ata-ST3320620A_x

 

beta11, 4GB ram, E8400, Asus P5E-VM D0, disk6 connected to SiI 3132 Controller, rest of the drives are on the motherboard, cache is only ide device

To those people like Daniel Boone who do not have an obvious "IDE mode" sata device table, it would be useful to know if either the media stream test or the telnet session test are reproducible scenarios. If you arent sure what I mean scan back through this thread.

 

Also please post your config details and also what disks show this behavior, what controllers control each disk.

 

In Daniels case he has two (ignoring cache) disk controllers so a simple set of elimination tests would be:

 

Streaming from disk 5 (mb), request disk 6 (sil3132) spin up disk5 (mb) stream stutters. Streaming from disk 6 (SIL3132), spin up disk 5 (mb), disk6 (SIL3132) stream doesn't stutter. Streaming from disk 5 (mb) spin up disk 4 (mb) disk 5(mb) doesn't stutter.

 

From those three test results we can be see the SIL3132 driver is using a command based call and pausing further disk operations pending completion of the disk request (I think this is a likely cause, we just need to identify the drivers which do this).

 

That MB is using a ICH9 southbridge which is command based so might well be the culprit (in the test above first one would pass, second two tests would stutter), if anyone has a ICH9 setup or a Asus P5E-VM D0 that doesnt have the issue please also let us know, likewise anyone with a SIL3132 setup. SIL3132 has been mentioned a couple of times.

 

 

Think it can get down to a 1100mhz Athlon XP?

Yep.  Mine is at 1000MHz.  And it doesn't even have any L2 cache.

The box is drawing ~30 Watts from the wall with 6 disks spun down.

 

That's about the same amount of power my Intel Core 2 Duo 2.4Ghz CPU with 4Gig DDR2 system is drawing with 3 disks spun down and 1 disk spun up. There are upgrade options out there if you ever decide to rid yourself of your low-performance headaches.

Thanks Kaygee,

 

I will test over the weekend.

I will say the media stall is not related to the SIL controller. The issue exsisted before its very recent installation.

 

For me I noticed my hardware player, a Popcorn hour, while watching a movie would stall for a few seconds whenever I copied, XP to unRaid, new content to the server. Eventually the PCH starts up again but in a fast forward state. This lasts for a few seconds until things return to normal.

 

It was noticed in all the prior betas. I never gave it much thought until it was described here. Hopefully its fixed in this release. 

 

Edit: Also I copy direct to the disk although I do have a User Shares.

Here are my devices;

 

pci-0000:00:11.0-scsi-0:0:0:0 (sda) ata-WDC_WD15EADS-00S2B0_WD-WCAVY0684188

pci-0000:00:11.0-scsi-1:0:0:0 (sdb) ata-ST31000340AS_9QJ1MB5Y

pci-0000:00:11.0-scsi-2:0:0:0 (sdc) ata-ST31000340AS_9QJ1GL28

pci-0000:00:11.0-scsi-3:0:0:0 (sdd) ata-WDC_WD10EAVS-32D7B1_WD-WCAU48166213

 

I should not get a studder looking at this list.

 

If I spin-up a drive using myMain a HD test video will play through the spin-up every time. I can spin-up and down drives as fast as the interface will go wiithout a studder.

 

If I try to access a directory or file from my windows explorer the same HD video will studder almost every time.

 

So, there must be more to the studdering video story than pairs of disks on a shared SATA controller. Maybe it's something to do with how Windows works?

 

Peter

 

Maybe it's something to do with how Windows works?

 

a few people (including myself) are getting this problem with OSX hosts, so it's likely unraid

my stuttering is also when accessing directories on disks that are spun down, I don't think it happens due to copying during media streaming. This is using OS X on a new Mac Mini and GB network.

I can confirm NOT having stuttering when streaming to one pc and accessing a spun down disk from another on both of my Win7 x64 machines. Trying to browse a spun down share and streaming a hd video to the same pc caused a momentary freeze in playback on both of them as well. The same results were found with both of my unraid servers. Both have 10 WD10EADS drives with #1 having ICH9R + 2 SAT2-MV8s, and #2 having ICH9R + 2 1430SAs. All drives appear to be on their own channels so this 100% appears to be the Windows network buffer issue, I tested on different HD files on the same disks to avoid streaming from the servers memory. Both VLC and MPC HC were used in case one was more aggressive with its playback buffer. With the prevelence of wireless networks and cheaper ram I'd expect more larger buffers options from players these days.

 

I went ahead and tested with 4.4.2 on the 2nd server, same results once again. With that version and motherboard in the past, ps3 streaming would get interrupted every time a spun down disk was accessed. It must have had something to do with the upnp software in combination with any other possible network access issues during a spin up. I only vaguely recall having interrupted streaming since switching to a XBMC box. But, because it prompts to resume at the stopped point, it wouldn't have been nearly as annoying.

 

Tomorrow I'll try browsing a spun down share with OS X and seeing if it interrupts XBMC playback. 2001's opening music didn't miss a beat when explorer spun up a file heavy directory the other day.