elkay14

xHCI is not enabled in the kernel provided with 4.7. I can create a custom kernel/bzroot for you that would enable it. PM me if you are interested.

Really nice build! Looks like it should be pretty quiet.

Post relevant syslog entries.

I don't know -- I didn't look at the changelog very deeply. I just thought I'd put it out there to see if it helped. That link you posted doesn't work.

That should make no difference whatsoever. I can only think you did not actually type none the first time and maybe put a space in there? In all my testing, I always just typed none without the quotes.

Weird -- your system does not have the 'messagebus' user or group set up on it. I would assume that you have upgraded this system from an earlier version of 4.X? You're going to have to add that useradd line to your go script or you'll have the same problem on reboot.

Linux kernel v3.2.0 is out. Here's a 3.2.0 unRAID kernel / bzroot to test on beta14: 3.2.0 files for beta14

What has IP 192.168.1.1 on your network? I assume it is a wireless router? These error messages are likely benign.

No. It means that for the purposes of unRAID, the Linux 3.x kernel is not stable enough yet. 3TB support in 4.7 is simply never going to happen as the necessary changes are essentially what v5.0 (beta) is. 3TB support under Fedora should be fine if it is a current enough version.

I've backported the bugfix from the 5.0beta series to the 4.7 kernel. Download the updated bzimage and bzroot files here. Changes: * Backport 5.0beta8 driver code (SG_IO spindown, unraid.c bugfix, etc) * Enable Marvell 9125/9128 SATA chipsets (Rocket 62X cards) * Powernow/Speedstep modules * MPT2SAS modules v03.100.03.00 (LSI SAS drivers)** * LZMA kernel and bzroot compression (much smaller bzimage and bzroot) PLEASE PLEASE PLEASE DO NOT use this with your production array yet. I have successfully done a replace disk procedure, but want many others to test this before I am comfortable with production use. Note that this kernel is optimized for AMD (what I have) but will work just fine with Intel. Once we are more certain this version is stable I will release an Intel optimized version. ----- ** Although spindown works with my reflashed M1015 card, my syslog fills quickly with harmless SCSI sense errors. It is my theory that the GUI is doing something the driver does not like, and as a result it appears the GUI does not seem to know that the drive is actually spun down -- it continually reissues spindown commands. Also, temperatures are not available via this driver as the GUI would need to be updated with the new temperature routines from the 5.0beta line. That is entirely in Lime Tech's hands.

Whoops; missed that.

http://www.newegg.com/Product/Product.aspx?Item=N82E16817341017&cm_sp=DailyDeal-_-17-341-017-_-Product +3.3V@25A, +5V@25A, +12V1@25A, +12V2@25A, [email protected], [email protected]

Getting the M1015 working with 4.7 is proving to be nearly futile. In the end, drive temps won't be supported (because that code is in the webgui) and I'm getting weird SCSI sense errors after the webgui issues spindown commands. It's not the spindown command itself, but the webgui is subsequently issuing a drive command of some type that the SCSI layer or even the card itself does not like. I do have the 4.7 RAID bug patched and will be looking for testing volunteers soon.

All of these are in the driver. The release notes say 'driver' and not 'WebGUI' for those items. The rebuild bug is most likely in unraid.c and not md.c; I'm still digging on that one. I have a new working md.c that does spindown/spinup on my test box with the SG_IO code. There's only a PATA drive in that box -- I will be testing with an M1015 in another server shortly.

I don't see how it could since none of the drivers are included: # CONFIG_MEGARAID_NEWGEN is not set # CONFIG_MEGARAID_LEGACY is not set # CONFIG_MEGARAID_SAS is not set # CONFIG_SCSI_MPT2SAS is not set That is from the .config for 4.7 root@Tower:/lib/modules/2.6.32.9-unRAID/kernel/drivers/scsi# find . . ./3w-9xxx.ko ./scsi_wait_scan.ko ./libsas ./libsas/libsas.ko ./arcmsr ./arcmsr/arcmsr.ko ./aacraid ./aacraid/aacraid.ko ./3w-xxxx.ko ./scsi_transport_sas.ko I have a 4.7 kernel running on my test box with the SG_IO interface spindown code and the driver for the M1015. I'll be testing it tonight.

It cannot be done. unRAID replaces the md (multidevice) kernel modules and both dm-crypt and devicemapper are under md. The kernel options are missing when unraid md code is installed into the kernel source tree.

Days to a couple weeks. I'm in the middle of learning the code and determining what needs to get ported back to 4.7. Definitely not months.

Mainly those two things: Spindown support for SAS/SCSI disks, RAID stripe bug. There are a couple of other bugfixes mentioned in the release notes that I am in the process of determining if they are already included in 4.7. It's a little tough considering the driver changed quite a bit with the introduction of 3TB support (which cannot be backported since a lot of the changes are also in the WebGUI.) Changes from 5.0-beta2 to 5.0-beta3 ----------------------------------- Bug Fixes: - driver: fix rare unraid driver kernel oops (protect md cmd/status interface with semaphore). This semaphore fix for the /proc interface is already in 4.7. Changes from 5.0-beta3 to 5.0-beta4 ----------------------------------- Bug Fixes: - driver: eliminate code in driver to get disk model/serial/size, instead use new i/f to pass slot info to driver. - driver: preliminary changes to support large (>2TB) hard drives [various 32-bit vars changed to 64-bit]. These cannot be backported. Changes from 5.0-beta5a to 5.0-beta5b ------------------------------------- Bug Fixes: - driver: fix detecting UPGRADE_DISK case. I've got this one included now. Changes from 5.0-beta5b to 5.0-beta6 ------------------------------------ Bug Fixes: - driver: fixed issue where super.dat file was being truncated. I haven't figured out if this one is included or not. Changes from 5.0-beta6b to 5.0-beta7 ------------------------------------ - driver: added SCSI/SAS device spin up/down support via SG_IO. I am working on this one; it shouldn't be too hard. Changes from 5.0-beta7 to 5.0-beta8 ----------------------------------- - driver: fixed bug causing parity errors with right combination of writes and parity-sync/data-rebuild I am working on this one; more difficult to figure out what changed. Changes from 5.0-beta11 to 5.0-beta12 ------------------------------------- - driver: incorporate "explicit block device plugging" (http://lwn.net/Articles/438256/) No need to backport.

I am running the same setup with ESXi 4.1. I believe the M1015 was not supported until 5.0 Beta 8. I am running Beta 9 still as I don't fully trust the newer betas yet. That's right, the M1015 is not supported on 4.7. I am currently working on getting it supported through a new kernel and rootfs. Stay tuned.

Thanks!

I'm backporting some of the 5.0 driver features to 4.7 (spindown for SAS, etc) and I would like some of the older betas to look at driver code diffs. Anything beta7 and older is not available for download anymore. Does anyone have the following: beta7 beta6 or 6a beta5b beta4 beta3 Let me know if you have them and I can supply you ftp credentials to upload them.

I'm actually working on both of these right now!

I can confirm that reflashing an M1015 to IT mode works on a Biostar A760G M2+ (one of the recommended motherboards for unRAID budget server, now EOL) I used the DOS tools -- I could not get the Linux ones to recognize a card.

Here's a 3.0.3 kernel for beta12a: files

No.