kernel-packages team mailing list archive

[Axel Pospischil <apos@xxxxxx>]

** Attachment added: "20140101-BUG-1265309-ubuntu-bug-report.txt"
   https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1265309/+attachment/3938581/+files/20140101-BUG-1265309-ubuntu-bug-report.txt

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux in Ubuntu.
https://bugs.launchpad.net/bugs/1265309

Title:
  [Lenovo W510 - SSD Samsung EVO] Sudden Read-Only Filesystems

Status in “linux” package in Ubuntu:
  Incomplete

Bug description:
  After a while of running the system, the SSD goes into a readonly file system mode. I am running Ubuntu 12.04 LTS linux-image-generic-lts-saucy, lvm2 in a version that supports trim and are encrypted on three machines, with the only one affected is 3.:
  1. An intel xeon with an asus motherboard, 32GB RAM, SSD (SAMSUNG older model)
  2. An Lenovo Thinkpad 201s i7, 8GB RAM, SSD (Crucial)
  3. An Lenovo Thinkpad W510 i7 720, 8GB RAM, SSD (Samsung EVO)

  ii lvm2 2.02.95-4ubuntu1.1~p Linux Logical Volume Manager

  /etc/default/grub: GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"

  /etc/modprobe.d/options: EMPTY

  The #3 (W510) randomly freezes with the following errors:

  > dmesg | grep ata1
  ata1: SATA max UDMA/133 abar m2048@0xf2627000 port 0xf2627100 irq 54
  ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
  ata1.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
  ata1.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out
  ata1.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
  ata1.00: ATA-9: Samsung SSD 840 EVO 500GB, EXT0BB0Q, max UDMA/133
  ata1.00: 976773168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
  ata1.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
  ata1.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out
  ata1.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
  ata1.00: configured for UDMA/133
  [...]

  These errors occur up to 5-10 times, then the SSD is set to readonly
  state like this:

  > demsg | grep ata1
  ata1: EH complete
  ata1: limiting SATA link speed to 1.5 Gbps
  ata1.00: exception Emask 0x52 SAct 0x1 SErr 0x1a80d00 action 0x6 frozen
  ata1.00: irq_stat 0x08000000, interface fatal error
  ata1: SError: { UnrecovData Proto HostInt 10B8B BadCRC LinkSeq TrStaTrns }
  ata1.00: failed command: READ FPDMA QUEUED
  ata1.00: cmd 60/08:00:d8:b9:27/00:00:05:00:00/40 tag 0 ncq 4096 in
  ata1.00: status: { DRDY }
  ata1: hard resetting link
  ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)

  ---------------------------------

  i am now trying the following settings:

  /etc/default/grub: GRUB_CMDLINE_LINUX_DEFAULT="quiet splash
  libata.force=noncq"

  /etc/modprobe.d/options: options libata noacpi=1

  /etc/default/tlp (which was on standard setting until now)

  # Hard disk advanced power management level: 1(max saving)..254(off)
  # Levels 1..127 may spin down the disk.
  # Separate values for multiple devices with spaces.
  DISK_APM_LEVEL_ON_AC="254 254"
  DISK_APM_LEVEL_ON_BAT="254 254"

  # SATA aggressive link power management (ALPM):
  # min_power/medium_power/max_performance
  SATA_LINKPWR_ON_AC=max_performance
  SATA_LINKPWR_ON_BAT=max_performance

  I will write ... after further testing.

  Please note: system #1. and #2. are running without any problems!

  ---------------------------------------------------
  AFTER 2 DAYS
  ---------------------------------------------------

  Unfortunetely with the settings from above i was not lucky this
  morning. When I turned on the W510 notebook (#3.) it stuck like
  before. I really think, this is system specific for this laptop,
  because the other computers with itdentical software are running
  flawlessly ... 24/7 (via suspend / resume).

  I also have to add, that all disks run on lvm based cryptsetup!

  >  cat /var/log/syslog.1 | grep ata1 | cut -d "[" -f2

      1.419360] ata1: SATA max UDMA/133 abar m2048@0xf2627000 port 0xf2627100 irq 53
      1.737774] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
      1.741004] ata1.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
      1.741014] ata1.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered out
      1.741020] ata1.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
      1.741298] ata1.00: ATA-9: Samsung SSD 840 EVO 500GB, EXT0BB0Q, max UDMA/133
      1.741303] ata1.00: 976773168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
      1.742899] ata1.00: ACPI cmd ef/02:00:00:00:00:a0 (SET FEATURES) succeeded
      1.742907] ata1.00: ACPI cmd f5/00:00:00:00:00:a0 (SECURITY FREEZE LOCK) filtered ou
      1.742913] ata1.00: ACPI cmd ef/10:03:00:00:00:a0 (SET FEATURES) filtered out
      1.743209] ata1.00: configured for UDMA/133

  -----------------------------------------------
  AFTER 5 DAYS
  _____________________________

  I changed my fstab and removed the "defaults"-entry:
  Probably this was the problem and the "defaults"-entry puts some mount options, that interfere with the ssd.

  fatab-NEW: /dev/mapper/vg--myvg-root / ext4 noatime,errors=remount-ro 0 1
  fstab-OLD: /dev/mapper/vg--myvg-root / ext4 defaults,noatime,errors=remount-ro 0 1

  I left the libata option in modprobe and kernel boot options:
  > cat /etc/modprobe.d/options
  options libata noacpi=1

  > cat /etc/default/grub
  GRUB_CMDLINE_LINUX_DEFAULT="quiet splash libata.force=noncq libata.noacpi=1"

  Since two days the system is running without any filesystem-freezes.

  I am attaching an ubuntu bug-report which can probably help to solve and analyse the problem.
  The system is up to date as time of writing and all necessary updates are applied.

  ----------------------------------------------
  AFTER 6 DAYS
  _____________________________

  In my last comment I hoped, the problem was solved
  (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1063354/comments/123).
  Unfortunately it is not. Yeterday morning my filesystem was readonly
  again.

  So I am going to

  - remove now all noatime, all trim-related entries from fstab, lvm.conf and cryptsetup
  - I am reverting to the presice version of lvm 2
  - I am setting back all acpi settings for libata in modprobe.d/options and default/grub
  - Reverting the bios to F9 -> default settings

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1265309/+subscriptions