ESXI: usb 1-1: reset high-speed USB device number 2 using ehci-pci


Recommended Posts

Dan thanks for the info!

 

I have a couple of readers floating around here, hopefully one has a unique GUID. If not I'll try the FCR HS3.

 

Can anyone explain why this would actually work? I cant understand why a reader works vs a flash drive, especially after trying multiple different brands/sizes of flash drives.

 

One other thing I tried while troubleshooting this was to update the X9SCM-iif BIOS to Revision 2.2 - but it didnt help.

Here is the  BIOS change log for 2.0a to 2.2 on X9SCM-IIF(From SuperMicro)

    1. Update PXE OPEOM to 1.4.4

    2. Enhance B2 workaround

    3. Remove ME power SDR

    4. Update Intel RST to 11.6

    5. Fix SBIOS type#17 memory speed not correct issue

    6. Update CPU microcode

    7. Fix Memory speed not correct issue

    8. update UEFI security fix

    9. Change BIOS revision to 2.2

 

Also this is how the USB ports are split for the 2 USB controllers:

    USB0-7 to EHCI0

    USB8-13 to EHCI1

 

 

 

Link to comment

I found an old no-name usb reader(ESXi lists it as "Alcor Micro Mass Storage Device") and it is working without the resets in ESI 5.5 with UnRaid 6.0 Final.

I am booting from VMDK with just the config folder on the reader.

 

I'm glad using the reader worked, but would be nice to be able to get flash drives working again.

 

 

To Limetech: The new key replacement feature worked very smooth!

 

 

 

 

Link to comment

I'm still seeing some issues regarding this (with a card reader that is supported by unraid, and a generic microsd card).

 

To recreate, I simply reboot, and navigate the web-ui until it hangs (basically all menus), it eventually does.  And then I get something along the following in my syslog:

 

Jun 26 23:10:06 phatstore kernel: usb 1-1: reset high-speed USB device number 2 using ehci-pci

 

It doesn't repeat like it did with a usb flash device, and I don't know if a better microsd card will resolve it, but just thought I'd post as a heads up.

 

 

 

Link to comment

ok, i spent much time on this issue, and there are my current observations:

fo me, it looks like it's flash dependent, cos i have two servers with very similar Lexar flash devices and both of them are resetting. on one server i have some sundisk flash too, and this one don't throw usb resets.

my tests are performed on Ubuntu VM's with various kernels, and i paste syslog fragments below from both working and non-working:

 

working one:

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.669283] scsi 3:0:0:0: Direct-Access    SanDisk  Cruzer Glide    1.26 PQ: 0 ANSI: 5

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.671284] sd 3:0:0:0: Attached scsi generic sg2 type 0

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.686596] sd 3:0:0:0: [sdb] 15633408 512-byte logical blocks: (8.00 GB/7.45 GiB)

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.700103] sd 3:0:0:0: [sdb] Write Protect is off

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.700106] sd 3:0:0:0: [sdb] Mode Sense: 43 00 00 00

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.712911] sd 3:0:0:0: [sdb] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA

Jun 25 14:31:02 i5TestUbuntu kernel: [    4.959287]  sdb: sdb1

Jun 25 14:31:02 i5TestUbuntu kernel: [    5.037076] sd 3:0:0:0: [sdb] Attached SCSI removable disk

 

nonworking:

Jun 25 13:22:46 ubuntu15 kernel: [    2.884010] scsi 2:0:0:0: Direct-Access    Lexar    USB Flash Drive  1100 PQ: 0 ANSI: 4

Jun 25 13:22:46 ubuntu15 kernel: [    2.884299] sd 2:0:0:0: Attached scsi generic sg2 type 0

Jun 25 13:22:46 ubuntu15 kernel: [    2.900210] sd 2:0:0:0: [sdb] 15659008 512-byte logical blocks: (8.01 GB/7.46 GiB)

Jun 25 13:22:46 ubuntu15 kernel: [    2.909449] sd 2:0:0:0: [sdb] Write Protect is off

Jun 25 13:22:46 ubuntu15 kernel: [    2.909455] sd 2:0:0:0: [sdb] Mode Sense: 43 00 00 00

Jun 25 13:22:46 ubuntu15 kernel: [    2.918925] sd 2:0:0:0: [sdb] No Caching mode page found

Jun 25 13:22:46 ubuntu15 kernel: [    2.919135] sd 2:0:0:0: [sdb] Assuming drive cache: write through

Jun 25 13:22:46 ubuntu15 kernel: [    2.978855]  sdb: sdb1

Jun 25 13:22:46 ubuntu15 kernel: [    3.015183] sd 2:0:0:0: [sdb] Attached SCSI removable disk

 

i bolded one message above which seem make sense..

may be some of you with better Linux knowledge can comment it out?

Link to comment

I am seeing:

    ...No Caching mode page found

    ...Assuming drive cache: write through

 

from my card reader(6.0) and SanDisk blade(6b15) which do not get the 'resets'.

and also on all multiple flash drives(which have the resets).

 

so it doesn't seem to be the cause.

I think it is just a regular check to see if the device has a useable cache(like a regular hard drive).

Link to comment

I am seeing:

    ...No Caching mode page found

    ...Assuming drive cache: write through

 

from my card reader(6.0) and SanDisk blade(6b15) which do not get the 'resets'.

and also on all multiple flash drives(which have the resets).

 

so it doesn't seem to be the cause.

I think it is just a regular check to see if the device has a useable cache(like a regular hard drive).

 

ok, but with those two lines you mention above unRAID v6b15 works just fine..

everything else i should test with this issue? i have no idea...

Link to comment

I found some old posts about the usb reset trouble, but it was on usb mouse+keyboard troubles.

  https://bbs.archlinux.org/viewtopic.php?id=91496

The fix for their issue was to enable kernel option 'CONFIG_USB_EHCI_TT_NEWSCHED'.

 

I don't know enough about linux to check if this could be a cause. Does anyone have a list or a command to get a list of all the UnRaid kernel options? I would like to compare the kernel options from UnRaid 6b15 to 6.0. I am thinking 1 of the default options may have flipped going to the 4.0 kernel.

 

Does anyone have a good(easy to understand) guide or website on compiling kernels? and a list of packages Unraid 6 uses?

 

more testing:

I tried formatting the flash drive as fat16, but no change in resets.

 

I also wanted to try running as USB1:

I tried disabling the ehci driver to force usb-1 on a flash drive in a 6.0 VM.

"modprobe -r ehci-pci" errors that it is builtin.

so i tried  'echo -n "{device id from lspci}" > /sys/bus/pci/drivers/ehci-pci/unbind'  (read it somewhere)

That does remove the usb 2  from 'lsusb' but then i don't have access to the flash drive. Is there another command i need to get access back as usb1? or is there a better way to test the drive as usb1?

 

 

Link to comment

Does anyone have a list or a command to get a list of all the UnRaid kernel options? I would like to compare the kernel options from UnRaid 6b15 to 6.0. I am thinking 1 of the default options may have flipped going to the 4.0 kernel.

 

on Ubuntu it should be:

cat "/boot/config-`uname -r`" | grep CONFIG_USB_EHCI

 

not sure it's the same for unRAID, cos /boot is your flash drive..

Link to comment

Just for $hit$ since I haven't messed with the final version 6 and my USB (which was previously seeing errors on all the RC3 and up), I decided to install 6.0.1 on my USB 3.0 San Disk Ultra Fit (configured with an xHCI USB Controller in VM settings) today along side a VMDK on my ESXi 5.5 server.  It's been up a half hour and I've had no USB reset errors since boot (I always get the message once or twice during boot even on my working 6b15 server).  Now it's not an official test since I have no other disks attached to this VM since my M1015 is passed through currently to my running 6b15 server but I'm not sure if adding disks would make any difference anyway with regard to this issue.

 

I'm tempted to try and upgrade my 6b15 server to 6.0.1 (worst case I just downgrade if it breaks it).  However that is a USB 2.0 flash drive connected to a physical USB 2.0 port and configured with an EHCI USB controller in my VM settings so it's not exactly a great test.

 

Have I read something somewhere about being able to change your USB flash drive through the webgui somewhere?

Link to comment

Add me to the list of people with the issue.  I did a fresh install of 6.0.1 last night on ESX6 with 2 hds and getting the really slow web gui and the reset errors.  I just haven't been able to start up yet as I am preclearing 2 8TB drives and I don't see a way to start without any assigned drives.

Link to comment

I have just done direct path io passthrough of the whole usb controller rather than adding a usb device and the issue has gone.  Web gui very fast no reset errors.  Only downside is plop will not boot had to use PlopKexec livecd now just trying to work out how to have it autoboot rather than needing to select to boot unraid.

 

Another thing to note is now the dashboard shows cpu usage as being 1-2% not the 99-100% it was before even tho top didn't show any cpu usage.

Link to comment

Just for $hit$ since I haven't messed with the final version 6 and my USB (which was previously seeing errors on all the RC3 and up), I decided to install 6.0.1 on my USB 3.0 San Disk Ultra Fit (configured with an xHCI USB Controller in VM settings) today along side a VMDK on my ESXi 5.5 server.  It's been up a half hour and I've had no USB reset errors since boot (I always get the message once or twice during boot even on my working 6b15 server).  Now it's not an official test since I have no other disks attached to this VM since my M1015 is passed through currently to my running 6b15 server but I'm not sure if adding disks would make any difference anyway with regard to this issue.

 

I'm tempted to try and upgrade my 6b15 server to 6.0.1 (worst case I just downgrade if it breaks it).  However that is a USB 2.0 flash drive connected to a physical USB 2.0 port and configured with an EHCI USB controller in my VM settings so it's not exactly a great test.

 

Have I read something somewhere about being able to change your USB flash drive through the webgui somewhere?

 

as i wrote above, my Sundisk USB2 flash works without reset errors for me too..

that's really weird if it's flash manufacturer dependent... 

Link to comment

so, had some more time for testing today, some new observations:

- since i added separate UPS to my test server, i plugged it in usb too, and added both unRAID flash and UPS to my Ubuntu test VM. and surprise, no more usb resets, leaved it for and hour, still no usb resets. then i restarted my Ubuntu VM with unRAID usb flash only and resets are back..

- next i tried unRAID v6 final VM with both usb's - booted and noticed resets started happening. then i setup UPS daemon, and voila, no more usb resets.. leaved it for an hour and half and no single usb reset.

 

Have anyone some comments and/or ideas what to try next? :)

Link to comment

I solved the problem by passing the usb controller on my X9SCM directly through to the unRaid vm.  I boot ESXi 5.5 U2 from the usb port on the mb so make sure you pt controller #1 (the ones on the rear I/O).  USB Controller #2 is for the internal ports (mb and mb header).

 

UnRaid boots from a vmdk image, so the USB only supplies the config folder.

 

Thanks,

Rich

Link to comment

I solved the problem by passing the usb controller on my X9SCM directly through to the unRaid vm.  I boot ESXi 5.5 U2 from the usb port on the mb so make sure you pt controller #1 (the ones on the rear I/O).  USB Controller #2 is for the internal ports (mb and mb header).

 

UnRaid boots from a vmdk image, so the USB only supplies the config folder.

 

Thanks,

Rich

 

I did the same with my backup server.  It worked perfectly.

 

Since then I have upgraded my main server to 6.0.1 (identical hardware).  Everything is now working without issue.  This is a good workaround until the cause of this issue is found and corrected.

Link to comment

Would it be possible to diff compare the last reported working version (b15?) to the current release? If it is isolated to to a change, the kernel as some suspect, would that allow for review by Tom and others the best way to fix.

 

I'm still at v5 and planning to stay there until this is resolved.

Link to comment

Would it be possible to diff compare the last reported working version (b15?) to the current release? If it is isolated to to a change, the kernel as some suspect, would that allow for review by Tom and others the best way to fix.

 

I'm still at v5 and planning to stay there until this is resolved.

 

Beta 16 had a kernel update - http://lime-technology.com/wiki/index.php/Release_Notes#Summaries_of_changes_from_6.0-beta15_to_6.0-rc3

Link to comment

Having to passthrough my USB controller is killing me because it renders all the rest of my USB ports useless (all my USB ports are on a single controller).  I hope this "bug" gets fixed soon.

 

Unless it gets fixed inadvertently, I doubt it will.  Limetech has stated that they do not support running unRAID as a guest.

Link to comment

Having to passthrough my USB controller is killing me because it renders all the rest of my USB ports useless (all my USB ports are on a single controller).  I hope this "bug" gets fixed soon.

 

Unless it gets fixed inadvertently, I doubt it will.  Limetech has stated that they do not support running unRAID as a guest.

 

Yea I know.  I guess I'm just hoping someone else running as a guest "stumbles" across a fix like you said.

Link to comment

Very likely, as it was broken in a kernel update, it will be fixed in a kernel update.  Somewhere out there is a user who may not have even heard of unRAID, but is suffering from the same problem.  They will report it, and diffs will be checked, and a patch will be reversed or corrected.

Link to comment

Very likely, as it was broken in a kernel update, it will be fixed in a kernel update.  Somewhere out there is a user who may not have even heard of unRAID, but is suffering from the same problem.  They will report it, and diffs will be checked, and a patch will be reversed or corrected.

 

Yes, may be :) but then you have to wait somewhere till the end of the year when more Linux distros(Ubuntu for example) start shipping kernel 4.

Link to comment

Very likely, as it was broken in a kernel update, it will be fixed in a kernel update.  Somewhere out there is a user who may not have even heard of unRAID, but is suffering from the same problem.  They will report it, and diffs will be checked, and a patch will be reversed or corrected.

 

Yes, may be :) but then you have to wait somewhere till the end of the year when more Linux distros(Ubuntu for example) start shipping kernel 4.

 

New 6.1 RC has new kernel... no time to update yet, but I'm curious if it fixes our issues.

Link to comment

I'm experiencing this problem as well.  I'm running Unraid 6 as a guest on ESXI 5.1 (and a supermicro X9SCM motherboard with SASLP-MV8 raid controllers).

However, I'm unable to pass through the USB controller -- when I set the value to pass it through, it doesn't 'stick' through the reboot.

 

Any one know if an upgrade to the latest ESXI would correct this?

 

 

Edit: I've upgraded to 6.1rc and am still seeing this same issue:

 

Jul 18 23:42:22 Tower kernel: usb 1-1: reset high-speed USB device number 2 using ehci-pci

Jul 18 23:42:22 Tower kernel: sd 3:0:0:0: [sda] 121307136 512-byte logical blocks: (62.1 GB/57.8 GiB)

Capture2.PNG.211a3f033d06d2da3e1c7cdbf4ae47c1.PNG

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.