3

I have a server with a RAID controller which connects to a JBOD over SAS. JBOD is populated with 16 SAS-HDDs of 3TB (8 of one make and 8 of other, same rpm). I have configured 3 logical drives with RAID-5, each having 5 physical HDDs (+1 as hot-spare).

Now to test it out, i run the following script:

for i in 1 10 50 100 1000
 do for j in a b c
  do dd if=/dev/zero of=/dev/sd$j bs=1G count=$i
 done
done

Everytime i run this script, it runs fine for the count=1,10 and 50.

But with 100G, writes on virtual drives crash randomly. Meaning, sometimes write on /dev/sda gets completed without and error but crashes for /dev/sdb. Sometimes writes on /dev/sda and /dev/sdb complete but fails on /dev/sdc.

I suspect that my RAID card might be faulty because i have already tested my harddisks individually by attaching them directly to the server and running "dd" on full 3TB.

What do you guys suggest?

EDIT:

Server motherboard make/model: SuperMicro X8DTH-6F
RAID controller: LSI MegaRaid SAS 9285-8e with BBU
JBOD: SuperMicro JBOD SC836E26-R1200B
SAS HDDs: 3.5", 6Gbps SAS, 3TB, 7200rpm ( 8x - Seagate ST3000NM0023, 8x - Hitachi Ultrastar 7K3000
OS: Scientific Linux 6.3

JBOD is connected to raid controller through a 6Gbps SAS cable.

EDIT 2: Here is the text from /var/log/messages:

May  1 18:14:41 fileserver udevd[875]: worker [4083] unexpectedly returned with status 0x0100
May  1 18:14:41 fileserver udevd[875]: worker [4083] failed while handling '/devices/pci0000:00/0000:00:03.0/0000:08:00.0/host0/target0:2:0/0:2:0:0/block/sda'
May  1 18:14:43 fileserver kernel: megasas: Found FW in FAULT state, will reset adapter.
May  1 18:14:43 fileserver kernel: megaraid_sas: resetting fusion adapter.
May  1 18:15:17 fileserver kernel: INFO: task dd:4144 blocked for more than 120 seconds.
May  1 18:15:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:15:17 fileserver kernel: dd            D 000000000000000d     0  4144   4132 0x00000004
May  1 18:15:17 fileserver kernel: ffff88060c513bf8 0000000000000082 0000000000000000 ffffea0015a3a8c0
May  1 18:15:17 fileserver kernel: ffff88062d837938 ffff88062d837848 ffff88062ce06ea0 ffff88062ce06ea0
May  1 18:15:17 fileserver kernel: ffff8806133385f8 ffff88060c513fd8 000000000000fb88 ffff8806133385f8
May  1 18:15:17 fileserver kernel: Call Trace:
May  1 18:15:17 fileserver kernel: [<ffffffff811140c0>] ? sync_page+0x0/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff814fdfc3>] io_schedule+0x73/0xc0
May  1 18:15:17 fileserver kernel: [<ffffffff811140fd>] sync_page+0x3d/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff814fe97f>] __wait_on_bit+0x5f/0x90
May  1 18:15:17 fileserver kernel: [<ffffffff81114333>] wait_on_page_bit+0x73/0x80
May  1 18:15:17 fileserver kernel: [<ffffffff81092110>] ? wake_bit_function+0x0/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff8112a835>] ? pagevec_lookup_tag+0x25/0x40
May  1 18:15:17 fileserver kernel: [<ffffffff811147ab>] wait_on_page_writeback_range+0xfb/0x190
May  1 18:15:17 fileserver kernel: [<ffffffff8111486f>] filemap_fdatawait+0x2f/0x40
May  1 18:15:17 fileserver kernel: [<ffffffff81114e94>] filemap_write_and_wait+0x44/0x60
May  1 18:15:17 fileserver kernel: [<ffffffff811b4634>] __sync_blockdev+0x24/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff811b4673>] sync_blockdev+0x13/0x20
May  1 18:15:17 fileserver kernel: [<ffffffff811b47f8>] __blkdev_put+0x178/0x1b0
May  1 18:15:17 fileserver kernel: [<ffffffff811b9403>] ? fsnotify+0x113/0x160
May  1 18:15:17 fileserver kernel: [<ffffffff811b4840>] blkdev_put+0x10/0x20
May  1 18:15:17 fileserver kernel: [<ffffffff811b4883>] blkdev_close+0x33/0x60
May  1 18:15:17 fileserver kernel: [<ffffffff8117c935>] __fput+0xf5/0x210
May  1 18:15:17 fileserver kernel: [<ffffffff8117ca75>] fput+0x25/0x30
May  1 18:15:17 fileserver kernel: [<ffffffff8117849d>] filp_close+0x5d/0x90
May  1 18:15:17 fileserver kernel: [<ffffffff81178575>] sys_close+0xa5/0x100
May  1 18:15:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:15:17 fileserver kernel: INFO: task scsi_id:4145 blocked for more than 120 seconds.
May  1 18:15:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:15:17 fileserver kernel: scsi_id       D 0000000000000013     0  4145      1 0x00000000
May  1 18:15:17 fileserver kernel: ffff880c1a145c18 0000000000000086 0000000000000000 ffff880c00000002
May  1 18:15:17 fileserver kernel: ffff880c1a145be8 00007f9200000066 00000000007fffff ffffffff81fc7990
May  1 18:15:17 fileserver kernel: ffff880c2ee3b098 ffff880c1a145fd8 000000000000fb88 ffff880c2ee3b098
May  1 18:15:17 fileserver kernel: Call Trace:
May  1 18:15:17 fileserver kernel: [<ffffffff814fee8e>] __mutex_lock_slowpath+0x13e/0x180
May  1 18:15:17 fileserver kernel: [<ffffffff8125ee50>] ? exact_match+0x0/0x10
May  1 18:15:17 fileserver kernel: [<ffffffff814fed2b>] mutex_lock+0x2b/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff811b4958>] __blkdev_get+0x68/0x3c0
May  1 18:15:17 fileserver kernel: [<ffffffff811b4cd0>] ? blkdev_open+0x0/0xc0
May  1 18:15:17 fileserver kernel: [<ffffffff811b4cc0>] blkdev_get+0x10/0x20
May  1 18:15:17 fileserver kernel: [<ffffffff811b4d41>] blkdev_open+0x71/0xc0
May  1 18:15:17 fileserver kernel: [<ffffffff8117888a>] __dentry_open+0x10a/0x360
May  1 18:15:17 fileserver kernel: [<ffffffff8121c262>] ? selinux_inode_permission+0x72/0xb0
May  1 18:15:17 fileserver kernel: [<ffffffff8121429f>] ? security_inode_permission+0x1f/0x30
May  1 18:15:17 fileserver kernel: [<ffffffff81178bf4>] nameidata_to_filp+0x54/0x70
May  1 18:15:17 fileserver kernel: [<ffffffff8118c100>] do_filp_open+0x6c0/0xd60
May  1 18:15:17 fileserver kernel: [<ffffffff8104452c>] ? __do_page_fault+0x1ec/0x480
May  1 18:15:17 fileserver kernel: [<ffffffff812718b1>] ? cpumask_any_but+0x31/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff811449a0>] ? unmap_region+0x110/0x130
May  1 18:15:17 fileserver kernel: [<ffffffff81198182>] ? alloc_fd+0x92/0x160
May  1 18:15:17 fileserver kernel: [<ffffffff81178639>] do_sys_open+0x69/0x140
May  1 18:15:17 fileserver kernel: [<ffffffff81178750>] sys_open+0x20/0x30
May  1 18:15:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:15:17 fileserver kernel: INFO: task fdisk:4176 blocked for more than 120 seconds.
May  1 18:15:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:15:17 fileserver kernel: fdisk         D 0000000000000013     0  4176   4147 0x00000004
May  1 18:15:17 fileserver kernel: ffff880c2f56dc18 0000000000000082 0000000000000000 ffff880c30802078
May  1 18:15:17 fileserver kernel: ffff880c2f56dbe8 ffff880c00000721 00000000007fffff ffffffff81fc8368
May  1 18:15:17 fileserver kernel: ffff880c1d6af058 ffff880c2f56dfd8 000000000000fb88 ffff880c1d6af058
May  1 18:15:17 fileserver kernel: Call Trace:
May  1 18:15:17 fileserver kernel: [<ffffffff814fee8e>] __mutex_lock_slowpath+0x13e/0x180
May  1 18:15:17 fileserver kernel: [<ffffffff8125ee50>] ? exact_match+0x0/0x10
May  1 18:15:17 fileserver kernel: [<ffffffff814fed2b>] mutex_lock+0x2b/0x50
May  1 18:15:17 fileserver kernel: [<ffffffff811b4958>] __blkdev_get+0x68/0x3c0
May  1 18:15:17 fileserver kernel: [<ffffffff811b4cd0>] ? blkdev_open+0x0/0xc0
May  1 18:15:17 fileserver kernel: [<ffffffff811b4cc0>] blkdev_get+0x10/0x20
May  1 18:15:17 fileserver kernel: [<ffffffff811b4d41>] blkdev_open+0x71/0xc0
May  1 18:15:17 fileserver kernel: [<ffffffff8117888a>] __dentry_open+0x10a/0x360
May  1 18:15:17 fileserver kernel: [<ffffffff8121c262>] ? selinux_inode_permission+0x72/0xb0
May  1 18:15:17 fileserver kernel: [<ffffffff8121429f>] ? security_inode_permission+0x1f/0x30
May  1 18:15:17 fileserver kernel: [<ffffffff81178bf4>] nameidata_to_filp+0x54/0x70
May  1 18:15:17 fileserver kernel: [<ffffffff8118c100>] do_filp_open+0x6c0/0xd60
May  1 18:15:17 fileserver kernel: [<ffffffff8104452c>] ? __do_page_fault+0x1ec/0x480
May  1 18:15:17 fileserver kernel: [<ffffffff811e06c5>] ? pde_users_dec+0x25/0x60
May  1 18:15:17 fileserver kernel: [<ffffffff81198182>] ? alloc_fd+0x92/0x160
May  1 18:15:17 fileserver kernel: [<ffffffff81178639>] do_sys_open+0x69/0x140
May  1 18:15:17 fileserver kernel: [<ffffffff81178750>] sys_open+0x20/0x30
May  1 18:15:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:16:27 fileserver kernel: megaraid_sas: Diag reset adapter never cleared!
May  1 18:17:17 fileserver kernel: INFO: task dd:4144 blocked for more than 120 seconds.
May  1 18:17:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:17:17 fileserver kernel: dd            D 000000000000000d     0  4144   4132 0x00000004
May  1 18:17:17 fileserver kernel: ffff88060c513bf8 0000000000000082 0000000000000000 ffffea0015a3a8c0
May  1 18:17:17 fileserver kernel: ffff88062d837938 ffff88062d837848 ffff88062ce06ea0 ffff88062ce06ea0
May  1 18:17:17 fileserver kernel: ffff8806133385f8 ffff88060c513fd8 000000000000fb88 ffff8806133385f8
May  1 18:17:17 fileserver kernel: Call Trace:
May  1 18:17:17 fileserver kernel: [<ffffffff811140c0>] ? sync_page+0x0/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff814fdfc3>] io_schedule+0x73/0xc0
May  1 18:17:17 fileserver kernel: [<ffffffff811140fd>] sync_page+0x3d/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff814fe97f>] __wait_on_bit+0x5f/0x90
May  1 18:17:17 fileserver kernel: [<ffffffff81114333>] wait_on_page_bit+0x73/0x80
May  1 18:17:17 fileserver kernel: [<ffffffff81092110>] ? wake_bit_function+0x0/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff8112a835>] ? pagevec_lookup_tag+0x25/0x40
May  1 18:17:17 fileserver kernel: [<ffffffff811147ab>] wait_on_page_writeback_range+0xfb/0x190
May  1 18:17:17 fileserver kernel: [<ffffffff8111486f>] filemap_fdatawait+0x2f/0x40
May  1 18:17:17 fileserver kernel: [<ffffffff81114e94>] filemap_write_and_wait+0x44/0x60
May  1 18:17:17 fileserver kernel: [<ffffffff811b4634>] __sync_blockdev+0x24/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff811b4673>] sync_blockdev+0x13/0x20
May  1 18:17:17 fileserver kernel: [<ffffffff811b47f8>] __blkdev_put+0x178/0x1b0
May  1 18:17:17 fileserver kernel: [<ffffffff811b9403>] ? fsnotify+0x113/0x160
May  1 18:17:17 fileserver kernel: [<ffffffff811b4840>] blkdev_put+0x10/0x20
May  1 18:17:17 fileserver kernel: [<ffffffff811b4883>] blkdev_close+0x33/0x60
May  1 18:17:17 fileserver kernel: [<ffffffff8117c935>] __fput+0xf5/0x210
May  1 18:17:17 fileserver kernel: [<ffffffff8117ca75>] fput+0x25/0x30
May  1 18:17:17 fileserver kernel: [<ffffffff8117849d>] filp_close+0x5d/0x90
May  1 18:17:17 fileserver kernel: [<ffffffff81178575>] sys_close+0xa5/0x100
May  1 18:17:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:17:17 fileserver kernel: INFO: task scsi_id:4145 blocked for more than 120 seconds.
May  1 18:17:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:17:17 fileserver kernel: scsi_id       D 0000000000000013     0  4145      1 0x00000000
May  1 18:17:17 fileserver kernel: ffff880c1a145c18 0000000000000086 0000000000000000 ffff880c00000002
May  1 18:17:17 fileserver kernel: ffff880c1a145be8 00007f9200000066 00000000007fffff ffffffff81fc7990
May  1 18:17:17 fileserver kernel: ffff880c2ee3b098 ffff880c1a145fd8 000000000000fb88 ffff880c2ee3b098
May  1 18:17:17 fileserver kernel: Call Trace:
May  1 18:17:17 fileserver kernel: [<ffffffff814fee8e>] __mutex_lock_slowpath+0x13e/0x180
May  1 18:17:17 fileserver kernel: [<ffffffff8125ee50>] ? exact_match+0x0/0x10
May  1 18:17:17 fileserver kernel: [<ffffffff814fed2b>] mutex_lock+0x2b/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff811b4958>] __blkdev_get+0x68/0x3c0
May  1 18:17:17 fileserver kernel: [<ffffffff811b4cd0>] ? blkdev_open+0x0/0xc0
May  1 18:17:17 fileserver kernel: [<ffffffff811b4cc0>] blkdev_get+0x10/0x20
May  1 18:17:17 fileserver kernel: [<ffffffff811b4d41>] blkdev_open+0x71/0xc0
May  1 18:17:17 fileserver kernel: [<ffffffff8117888a>] __dentry_open+0x10a/0x360
May  1 18:17:17 fileserver kernel: [<ffffffff8121c262>] ? selinux_inode_permission+0x72/0xb0
May  1 18:17:17 fileserver kernel: [<ffffffff8121429f>] ? security_inode_permission+0x1f/0x30
May  1 18:17:17 fileserver kernel: [<ffffffff81178bf4>] nameidata_to_filp+0x54/0x70
May  1 18:17:17 fileserver kernel: [<ffffffff8118c100>] do_filp_open+0x6c0/0xd60
May  1 18:17:17 fileserver kernel: [<ffffffff8104452c>] ? __do_page_fault+0x1ec/0x480
May  1 18:17:17 fileserver kernel: [<ffffffff812718b1>] ? cpumask_any_but+0x31/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff811449a0>] ? unmap_region+0x110/0x130
May  1 18:17:17 fileserver kernel: [<ffffffff81198182>] ? alloc_fd+0x92/0x160
May  1 18:17:17 fileserver kernel: [<ffffffff81178639>] do_sys_open+0x69/0x140
May  1 18:17:17 fileserver kernel: [<ffffffff81178750>] sys_open+0x20/0x30
May  1 18:17:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:17:17 fileserver kernel: INFO: task fdisk:4176 blocked for more than 120 seconds.
May  1 18:17:17 fileserver kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May  1 18:17:17 fileserver kernel: fdisk         D 0000000000000013     0  4176   4147 0x00000004
May  1 18:17:17 fileserver kernel: ffff880c2f56dc18 0000000000000082 0000000000000000 ffff880c30802078
May  1 18:17:17 fileserver kernel: ffff880c2f56dbe8 ffff880c00000721 00000000007fffff ffffffff81fc8368
May  1 18:17:17 fileserver kernel: ffff880c1d6af058 ffff880c2f56dfd8 000000000000fb88 ffff880c1d6af058
May  1 18:17:17 fileserver kernel: Call Trace:
May  1 18:17:17 fileserver kernel: [<ffffffff814fee8e>] __mutex_lock_slowpath+0x13e/0x180
May  1 18:17:17 fileserver kernel: [<ffffffff8125ee50>] ? exact_match+0x0/0x10
May  1 18:17:17 fileserver kernel: [<ffffffff814fed2b>] mutex_lock+0x2b/0x50
May  1 18:17:17 fileserver kernel: [<ffffffff811b4958>] __blkdev_get+0x68/0x3c0
May  1 18:17:17 fileserver kernel: [<ffffffff811b4cd0>] ? blkdev_open+0x0/0xc0
May  1 18:17:17 fileserver kernel: [<ffffffff811b4cc0>] blkdev_get+0x10/0x20
May  1 18:17:17 fileserver kernel: [<ffffffff811b4d41>] blkdev_open+0x71/0xc0
May  1 18:17:17 fileserver kernel: [<ffffffff8117888a>] __dentry_open+0x10a/0x360
May  1 18:17:17 fileserver kernel: [<ffffffff8121c262>] ? selinux_inode_permission+0x72/0xb0
May  1 18:17:17 fileserver kernel: [<ffffffff8121429f>] ? security_inode_permission+0x1f/0x30
May  1 18:17:17 fileserver kernel: [<ffffffff81178bf4>] nameidata_to_filp+0x54/0x70
May  1 18:17:17 fileserver kernel: [<ffffffff8118c100>] do_filp_open+0x6c0/0xd60
May  1 18:17:17 fileserver kernel: [<ffffffff8104452c>] ? __do_page_fault+0x1ec/0x480
May  1 18:17:17 fileserver kernel: [<ffffffff811e06c5>] ? pde_users_dec+0x25/0x60
May  1 18:17:17 fileserver kernel: [<ffffffff81198182>] ? alloc_fd+0x92/0x160
May  1 18:17:17 fileserver kernel: [<ffffffff81178639>] do_sys_open+0x69/0x140
May  1 18:17:17 fileserver kernel: [<ffffffff81178750>] sys_open+0x20/0x30
May  1 18:17:17 fileserver kernel: [<ffffffff8100b0f2>] system_call_fastpath+0x16/0x1b
May  1 18:18:11 fileserver kernel: megaraid_sas: Diag reset adapter never cleared!
...
...

I can't make anything of it, its a cry for help!

7
  • 1
    Server make/model? Enclosure make/model? Disk makes/models? RAID controller make/model? Operating system/version/distribution? Server specifications? Cabling arrangement?
    – ewwhite
    Apr 30, 2014 at 13:37
  • 1
    "but crashes for /dev/sdb" Stop making up terms. Unless dd itself actually crashed. Show the errors from the console, dmesg and the RAID controller.
    – MikeyB
    Apr 30, 2014 at 14:12
  • Ah, Supermicro... Have you checked firmware of the related components?
    – ewwhite
    Apr 30, 2014 at 14:20
  • @MikeyB Sorry, TIL. There is no error on console, i'll add the output of dmesg shortly. BTW how do i check errors from RAID controller?
    – hrs
    Apr 30, 2014 at 15:24
  • @ewwhite i haven't updates any firmware yet, will do that.
    – hrs
    Apr 30, 2014 at 15:26

1 Answer 1

1

Turned out that Scientific Linux 6.3 on Supermicro servers has some issues with PCIe. A friend suggested to add the following two options to grub:

  1. pcie_aspm=off
  2. disable_msi=1

After booting with these options, everything started working fine.
Any thoughts?

2
  • 1
    Ropey hardware, by which I mean, hardware that doesn't precisely honour the spec, and a kernel that's under-tolerant of this imprecision. Happens a lot, especially with budget kit and/or consumer-grade parts. If you're confident in your findings, ditch either SuperMicro or Scientific Linux. Personally, I'd ditch the hardware manufacturer. And thanks for coming back and posting the answer you found; please accept it as soon as the system will let you do so.
    – MadHatter
    May 26, 2014 at 8:00
  • @MadHatter thanks, i'll try different manufacturer next time.
    – hrs
    May 26, 2014 at 10:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .