Skip to contentSkip to conversation

do_root_fs failing on flock command

DISCUSSION Jack Mitchell 3 participants 5 messages Aug 27, 2013

lore
Share & tools
lore

Jack Mitchell

author
#0 lore raw

I'm attempting to build core-image-minimal using oe-core + meta-beagleboard on an NFS mount. I'm running into the following error:

Quoted 32 lines

[jack@archAntec beaglebone]$ bitbake core-image-minimal
Loading cache: 100% |#######################################################################################################| ETA: 00:00:00 Loaded 1193 entries from dependency cache.

Build Configuration:

BB_VERSION        = "1.19.1"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "Arch-Linux"
TARGET_SYS        = "arm-oe-linux-gnueabi"
MACHINE           = "beaglebone"
DISTRO_VERSION    = "oe-core.0"
TUNE_FEATURES     = "armv7a vfp neon"
TARGET_FPU        = "vfp-neon"
meta              = "master:2ac7783e04f5e8e6005f967e1a6dd65d2fc6a19a"
common-bsp        = "master:f02fabd39cd49bf5932fa5f90e83084d49a8cbe2"

NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
ERROR: Logfile of failure stored in: /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520 Log data follows:

| DEBUG: Executing python function rootfs_process_ignore
| DEBUG: Python function rootfs_process_ignore finished
| DEBUG: Executing python function rootfs_runtime_mapping
| DEBUG: Python function rootfs_runtime_mapping finished
| DEBUG: Executing shell function do_rootfs
| flock: /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/deploy/ipk/Packages.flock: Bad file descriptor
| WARNING: /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/run.do_rootfs.19520:1 exit 65 from
|   flock $pkgdir/Packages.flock -c "opkg-make-index -r $pkgdir/Packages -p $pkgdir/Packages -m $pkgdir/"
| ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
Quoted 7 lines

ERROR: Task 7 (/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb, do_rootfs) failed with exit code '1'
NOTE: Tasks Summary: Attempted 1580 tasks of which 1579 didn't need to be rerun and 1 failed.
No currently running tasks (1579 of 1581)

Summary: 1 task failed:
/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb, do_rootfs
Summary: There was 1 ERROR message shown, returning a non-zero exit code.

Does anyone have any insight to why this may be failing. It mentions a bad file descriptor, so I assume it is related to building on an NFS mount.

Cheers,
Jack.

Signature · 5 lines
  Jack Mitchell ([email protected])
  Embedded Systems Engineer
  Cambrideshire, UK
  http://www.embed.me.uk
--
#1 lore raw
Quoted 2 lines · Jack Mitchell

Does anyone have any insight to why this may be failing. It mentions a bad file descriptor, so I assume it is related to building on an NFS mount.

There's no particular reason that flock() oughtn't to work on NFS. If you happened to be mounted with -o nolock then you wouldn't get the correct semantics versus lockers on other hosts, but you shouldn't be getting -EBADF under any circumstances that I can think of.

If you run flock by hand against a file on the same mount, does it work? If it fails, what does the output from strace look like?

p.

#3 lore raw
In reply to Jack Mitchell
Quoted 52 lines · Jack Mitchell

I'm attempting to build core-image-minimal using oe-core + meta-beagleboard on an NFS mount. I'm running into the following error:

Quoted 49 lines

[jack@archAntec beaglebone]$ bitbake core-image-minimal Loading cache: 100%

|#######################################################################################################| 
ETA:  00:00:00

Loaded 1193 entries from dependency cache.

Build Configuration:

BB_VERSION        = "1.19.1"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "Arch-Linux"
TARGET_SYS        = "arm-oe-linux-gnueabi"
MACHINE           = "beaglebone"
DISTRO_VERSION    = "oe-core.0"
TUNE_FEATURES     = "armv7a vfp neon"
TARGET_FPU        = "vfp-neon"
meta              = "master:2ac7783e04f5e8e6005f967e1a6dd65d2fc6a19a"
common-bsp        = "master:f02fabd39cd49bf5932fa5f90e83084d49a8cbe2"

NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
ERROR: Logfile of failure stored in:
/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520 Log data follows:

| DEBUG: Executing python function rootfs_process_ignore
| DEBUG: Python function rootfs_process_ignore finished
| DEBUG: Executing python function rootfs_runtime_mapping
| DEBUG: Python function rootfs_runtime_mapping finished
| DEBUG: Executing shell function do_rootfs
| flock:

/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/deploy/ipk/Packages.flock:
Bad file descriptor
| WARNING:
/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/run.do_rootfs.19520:1 exit 65 from
| flock $pkgdir/Packages.flock -c "opkg-make-index -r
$pkgdir/Packages -p $pkgdir/Packages -m $pkgdir/"
| ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
ERROR: Task 7
(/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb,
do_rootfs) failed with exit code '1'
NOTE: Tasks Summary: Attempted 1580 tasks of which 1579 didn't need to
be rerun and 1 failed.
No currently running tasks (1579 of 1581)

Quoted 7 lines
Quoted 4 lines

Summary: 1 task failed:
/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb, do_rootfs
Summary: There was 1 ERROR message shown, returning a non-zero exit code.

Does anyone have any insight to why this may be failing. It mentions a bad file descriptor, so I assume it is related to building on an NFS mount.

Correct, running builds on NFS isn't really recommended or supported (other than the downloads and sstate directories). I looks like the locking didn't work on the NFS mount and reading the man page for flock suggests that even if you do get NFS locking working, its still not recommended.

Cheers,

Richard

author
#4 lore raw
Quoted 67 lines · Richard Purdie
Quoted 52 lines · Jack Mitchell

I'm attempting to build core-image-minimal using oe-core + meta-beagleboard on an NFS mount. I'm running into the following error:

Quoted 49 lines

[jack@archAntec beaglebone]$ bitbake core-image-minimal Loading cache: 100%

|#######################################################################################################|
ETA:  00:00:00

Loaded 1193 entries from dependency cache.

Build Configuration:

BB_VERSION        = "1.19.1"
BUILD_SYS         = "x86_64-linux"
NATIVELSBSTRING   = "Arch-Linux"
TARGET_SYS        = "arm-oe-linux-gnueabi"
MACHINE           = "beaglebone"
DISTRO_VERSION    = "oe-core.0"
TUNE_FEATURES     = "armv7a vfp neon"
TARGET_FPU        = "vfp-neon"
meta              = "master:2ac7783e04f5e8e6005f967e1a6dd65d2fc6a19a"
common-bsp        = "master:f02fabd39cd49bf5932fa5f90e83084d49a8cbe2"

NOTE: Resolving any missing task queue dependencies
NOTE: Preparing runqueue
NOTE: Executing SetScene Tasks
NOTE: Executing RunQueue Tasks
ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
ERROR: Logfile of failure stored in:
/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520 Log data follows:

| DEBUG: Executing python function rootfs_process_ignore
| DEBUG: Python function rootfs_process_ignore finished
| DEBUG: Executing python function rootfs_runtime_mapping
| DEBUG: Python function rootfs_runtime_mapping finished
| DEBUG: Executing shell function do_rootfs
| flock:

/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/deploy/ipk/Packages.flock:
Bad file descriptor
| WARNING:
/mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/run.do_rootfs.19520:1 exit 65 from
| flock $pkgdir/Packages.flock -c "opkg-make-index -r
$pkgdir/Packages -p $pkgdir/Packages -m $pkgdir/"
| ERROR: Function failed: do_rootfs (log file is located at /mnt/mediaserver/Projects/oe-core/beaglebone/tmp-eglibc/work/beaglebone-oe-linux-gnueabi/core-image-minimal/1.0-r0/temp/log.do_rootfs.19520)
ERROR: Task 7
(/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb,
do_rootfs) failed with exit code '1'
NOTE: Tasks Summary: Attempted 1580 tasks of which 1579 didn't need to
be rerun and 1 failed.
No currently running tasks (1579 of 1581)

Quoted 7 lines
Quoted 4 lines

Summary: 1 task failed:
/mnt/mediaserver/Projects/oe-core/meta/recipes-core/images/core-image-minimal.bb, do_rootfs
Summary: There was 1 ERROR message shown, returning a non-zero exit code.

Does anyone have any insight to why this may be failing. It mentions a bad file descriptor, so I assume it is related to building on an NFS mount.

Correct, running builds on NFS isn't really recommended or supported (other than the downloads and sstate directories). I looks like the locking didn't work on the NFS mount and reading the man page for flock suggests that even if you do get NFS locking working, its still not recommended.

Quoted 4 lines

Cheers,

Richard

Indeed, I've just been doing some further research myself and it seems as though flock on NFS isn't recommended. It states using fnctl is a better alternative, so I'm looking to see if there is a way to do that natively in bash which could possibly replace the use of flock. I have done all the build on NFS and this is the only issue I have run into, so I feel it's probably worthwhile to see if there is a reasonable work around.

Cheers,

Signature · 5 lines
  Jack Mitchell ([email protected])
  Embedded Systems Engineer
  Cambrideshire, UK
  http://www.embed.me.uk
--
author
#2 lore raw
In reply to Phil Blundell
Quoted 9 lines · Phil Blundell
Quoted 2 lines · Jack Mitchell

Does anyone have any insight to why this may be failing. It mentions a bad file descriptor, so I assume it is related to building on an NFS mount.

There's no particular reason that flock() oughtn't to work on NFS. If you happened to be mounted with -o nolock then you wouldn't get the correct semantics versus lockers on other hosts, but you shouldn't be getting -EBADF under any circumstances that I can think of.

Quoted 6 lines

If you run flock by hand against a file on the same mount, does it work? If it fails, what does the output from strace look like?

p.

http://pastebin.com/LB7nZs1q

Signature · 5 lines
  Jack Mitchell ([email protected])
  Embedded Systems Engineer
  Cambrideshire, UK
  http://www.embed.me.uk
--