[Solved] Re: Beta 5.0 RC8a Hard Drive Red Ball Still

[daniel.boone]

Same issue I've had since the later betas. I load 4.7 and drive works fine. In fact it has worked fine for some time now. At this point I'm inclined to swap out the drive completely since I don't want to go backwards. The drive in question is a 3TB seagate to which I added a HPA so it would function with the mainstream release.

 

I spotted this in the log file right after spin up.

 

Oct 12 09:02:44 Tower kernel: mdcmd (119): spinup 0

Oct 12 09:02:44 Tower kernel: mdcmd (120): spinup 1

Oct 12 09:02:44 Tower kernel: mdcmd (121): spinup 2

Oct 12 09:02:44 Tower kernel: mdcmd (122): spinup 3

Oct 12 09:02:44 Tower kernel: mdcmd (123): spinup 4

Oct 12 09:02:44 Tower kernel: mdcmd (124): spinup 5

Oct 12 09:02:44 Tower kernel: mdcmd (125): spinup 6

Oct 12 09:02:44 Tower kernel: mdcmd (126): spinup 7

Oct 12 09:02:44 Tower kernel: mdcmd (127): spinup 8

Oct 12 09:02:44 Tower kernel: mdcmd (128): spinup 9

Oct 12 09:02:44 Tower kernel: mdcmd (129): spinup 10

Oct 12 09:02:44 Tower kernel: mdcmd (130): spinup 11

Oct 12 09:02:44 Tower kernel: mdcmd (131): spinup 12

Oct 12 09:02:44 Tower kernel: mdcmd (132): spinup 13

Oct 12 09:02:44 Tower kernel: mdcmd (133): spinup 14

Oct 12 09:02:56 Tower kernel: drivers/scsi/mvsas/mv_sas.c 1952:Release slot [0] tag[0], task [e6926780]:

Oct 12 09:02:56 Tower kernel: sas: sas_ata_task_done: SAS error 8a

Oct 12 09:02:56 Tower kernel: sd 5:0:0:0: [sdk] command ec61f480 timed out

Oct 12 09:02:57 Tower kernel: sas: Enter sas_scsi_recover_host busy: 1 failed: 1

 

 

I have 16 drives in a 4U case, a 750 watt power supply and a SuperMicro MB. Cache is used an installation point. All plugins are disabled but I am loading an updated version of samba. I did run a parity check right after the upgrade and got no errors when complete.  I had to zip log due to size limits.

 

Smart fails right now. I'll test and post results once I can reboot.

 

Cheers

DB

syslog-2012-10-12.zip

[bcbgboy13]

Your signature is incorrect as you do not have Supermicro SASLP-MV8 - you have LSI SAS2008 controller - perhaps cross-flashed IBM M1015.

 

And this controller has an older P11 firmware  (and BIOS):

 

Oct 10 09:54:07 Tower kernel: mpt2sas0: LSISAS2008: FWVersion(11.00.00.00), ChipRevision(0x03), BiosVersion(07.21.00.00)

 

The latest firmware is P14 and they have some fixes that may be applicable to you and "cure" that Seagate

 

Your motherboard BIOS is also older - however I do not use Supermicro and cannot comment on the usefulness of upgrading this one too (apparently there could be some problems if upgraded to the newer 2.xx - but I think this is a must for compatibility with LSI controllers)

[daniel.boone]

Your signature is incorrect as you do not have Supermicro SASLP-MV8 - you have LSI SAS2008 controller - perhaps cross-flashed IBM M1015.

 

And this controller has an older P11 firmware  (and BIOS):

 

Oct 10 09:54:07 Tower kernel: mpt2sas0: LSISAS2008: FWVersion(11.00.00.00), ChipRevision(0x03), BiosVersion(07.21.00.00)

 

The latest firmware is P14 and they have some fixes that may be applicable to you and "cure" that Seagate

 

Your motherboard BIOS is also older - however I do not use Supermicro and cannot comment on the usefulness of upgrading this one too (apparently there could be some problems if upgraded to the newer 2.xx - but I think this is a must for compatibility with LSI controllers)

 

My sig was pretty accurate until I added the IBM card and the latest RC candidate.

 

I do have the SuperMicro MV8 as well as an Adaptec and a re-flashed IBM card attached to that MB. I have 8 ports from my MV8 HBA, 4 from Adaptec and the rest from my MB to make for a total of 16 drives in this array counting the cache and parity.

 

At this time the IBM controller has no array drives attached to it. I just plugged it in in the hope the beta worked and I could add more drives. This card brings the tally to 24 ports on my system.  It's unlikely the IBM controller is related to this issue. The seagate drive redballs with or without the IBM card.

 

This exact system works without fault when I load 4.7. I load any beta including the the latest RC a redball happens on this one drive every time.

 

I do appreciate the info on the FW. I'll make sure to update once I have the drive issue sorted.

 

Thanks

DB

[daniel.boone]

Also wanted to point out the failure seems to happen most frequently after a successful parity check and the drives have spun down.

 

" Parity is Valid:.  Last parity check < 1 day ago with no sync errors. "

 

Edit: Smart report added

SmartReport.txt

[daniel.boone]

I may have traced the root cause to a "working" usb stick. Once I swapped with my backup things stabilized quickly.

 

I went ahead and swap the 3TB seagate parity with a 3TB hitachi anyway. I wasn't comfortable leaving it after so much trouble.

 

I ran the seagate through preclear after removing the HPA. Tested good so I added it as a blank to the array just to see if it hold up. Been about a week and no issues. I've only added one disc so far. If my luck continues after a full month I will add some real content to the disk.

 

I've also updated the FW on the M1015. That fixed me not seeing the hard drives connected to that HBA.