On 1/12/20 10:54 AM, Alan Feuerbacher wrote:
On 1/12/2020 9:00 AM, Bruce Dubbs wrote:
On 1/12/20 12:33 AM, Alan Feuerbacher wrote:
On 1/11/2020 11:19 PM, Bruce Dubbs wrote:
On 1/11/20 11:46 PM, Alan Feuerbacher wrote:
I'm building the development version of LFS systemd, and today
found that a few programs had been updated, so I'm building them
again.
e2fsprogs-1.45.5 built ok, but make check is still hanging after 5
hours. I'll let it run all night and see what happens. Any clues
as to the hanging?
Version 1.45.4 checked ok.
The checks took about seven minutes for me. Most tests output a
line like
f_h_badnode: hash directory with bad HTREE nodes: ok
What was the last test that passed for you?
-- Bruce
I'm not quite sure how to answer that. There are many pages of test
output. The last line before the hang is:
m_hugefile_slack: mke2fs create hugefile fs with slack: ok
The next lines should be
f_bbfile: bad blocks in files: ok
f_zero_group: fallback for damaged group descriptors: ok
j_short_trans: transaction nuking the bitmaps: ok
d_fallocate: fallocate sparse files and big files: ok
f_preen: preen shouldn't destroy backup superblocks: ok
r_move_itable: filesystem resize which requires moving the inode
table: ok
...
You can try deleting tests/f_bbfile and rerunning the tests.
You can also chech to see how much space you have free in the lfs
partition and how much memory you have free.
Other than that, I do not have any other ideas.
-- Bruce
After hanging for 13 hours, I Control-C'd it, but that didn't kill the
process. Apparently the Fedora host system had also hung, since when I
tried to do other normal things, various things were broken. I
couldn't even log out normally, and had to power down the computer.
After I powered back up, Fedora seemed ok, so I went back into the
chroot environment in my normal way, and tried rebuilding the previous
e2fsprogs-1.45.4 . Although it passed the tests the first time around
in Section 6 of the LFS book, this time it failed the tests (I've
saved logfiles of make and make check). So I have to conclude that
something in the later software installations (I've installed a great
many BLFS programs) has broken e2fsprogs. Comments?
Then I tried building e2fsprogs-1.45.5 again. This time, make check
ran without hanging, but ended with virtually the same test failure
messages as for 1.45.4 (I've saved logfiles of make and make check).
The last bit of failure message is:
##########
. . .
making check in tests
make[1]: Entering directory '/sources/e2fsprogs-1.45.5/build/tests'
Running e2fsprogs test suite...
/bin/cp ../../tests/mke2fs.conf.in mke2fs.conf
../../tests/scripts/gen-test-data > test_data.tmp
Creating test_one script...
f_illitable_flexbg: illegal inode table with FLEX_BG: ok
r_expand_full: expand a totally full filesystem: failed
f_imagic_fs: imagic filesystem with imagic inodes: ok
f_badsymlinks: corrupted symlinks: ok
m_64bit_flexbg: mkfs with 64bit and flex_bg: ok
u_mke2fs_opt_oddsize: e2undo with mke2fs -z and non-32k-aligned bdev
size: failed
f_baddir2: salvage last directory entry: ok
d_dumpe2fs_group_only: dumpe2fs group only mode: ok
../../tests/u_corrupt_blk_csum_force/script: line 22: / 1024 : syntax
error: operand expected (error token is "/ 1024 ")
make[1]: *** [Makefile:384: ../../tests/u_corrupt_blk_csum_force] Error
1
make[1]: Leaving directory '/sources/e2fsprogs-1.45.5/build/tests'
make: *** [Makefile:419: check-recursive] Error 1
##########
At this point I'm thinking of starting from scratch on a new hard
drive, but that's an awful lot of time and work, and I don't want to
do that until I understand what's happening with the above tests.
Alan
Good morning Alan,
What version of the kernel is in use on your host system? I noticed you
said that you'd completed some BLFS packages - if you reboot into LFS
and try the tests again, does it have the same problem? What you said
originally sounds like a kernel panic or OOPS.
On my development system (a 4-core Core i5-6600k Skylake machine), I had
no problems with the tests. I have almost all of BLFS installed,
including all of the packages in the Filesystems chapter (I have to for
Parted and friends for gvfs + udisks). My kernel version is 5.3.14.
u_mke2fs: e2undo with mke2fs: ok
f_encrypted_lpf: encrypted lost+found directory: ok
r_64bit_big_expand: : skipped (slow test)
r_resize_inode: filesystem resize with a resize_inode present: ok
f_resize_inode: e2fsck with resize_inode: ok
e_brel_bma: block relocation table using memory array implementation:
skipped
f_inlinedata_flags: check incorrect inline_data flags: ok
m_raid_opt: raid options: ok
m_minrootdir: create fs image from dir, then minimize it: ok
f_dirlink: directory hard links: ok
357 tests succeeded 0 tests failed
Creating test_script...
make[1]: Leaving directory '/sources/e2fsprogs-1.45.4/build/tests'
The tests took me about a 60 seconds to complete.
- Doug
--
http://lists.linuxfromscratch.org/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
Do not top post on this list.
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
http://en.wikipedia.org/wiki/Posting_style
