WSL: Busy hang running yum in `__env_region_extend`

  • A brief description I observe a busy hang when running yum to install some packages. gcore fails to produce a useful core but attaching once hung provides a stack.

  • Expected results yum installs specified packages in specified root

  • Actual results (with terminal output if applicable) yum hangs part way through - I’ve left this alone long enough for it to run up 11h of CPU time; it’s definitely not making progress.

  • Your Windows build number 15063.0

  • Steps / All commands required to reproduce the error from a brand new installation

# run as root

apt-get update && apt-get install -y yum

rm -fr /tmp/testroot && mkdir -p /tmp/testroot/etc/yum && echo "[main]
cachedir=/var/cache/yum
keepcache=1
debuglevel=1
logfile=/var/log/yum.log
exactarch=1
obsoletes=1

[photon-updates]
name=VMware Photon Linux 1.0(x86_64)
baseurl=https://dl.bintray.com/vmware/photon_updates_1.0_x86_64
gpgcheck=0
enabled=1

[photon]
name=VMware Photon Linux 1.0(x86_64)
baseurl=https://vmware.bintray.com/photon_release_1.0_x86_64/
gpgcheck=0
enabled=1
" > /tmp/testroot/etc/yum/yum.conf

/usr/bin/python /usr/bin/yum --installroot /tmp/testroot install filesystem coreutils linux-esx --nogpgcheck -y -v
  • Strace of the failing command Attached - obviously not the same invocation as the gdb output, but it is reliable on my system and has been almost 100% recreate for several months. strace.txt

  • Required packages and commands to install yum

gdb stack:

(gdb) where
#0  0x00007f061062c5fb in __env_region_extend () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#1  0x00007f061062c825 in __env_alloc () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#2  0x00007f061064dabb in __memp_alloc () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#3  0x00007f0610651874 in __memp_fget () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#4  0x00007f0610610f5d in __db_new () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#5  0x00007f0610615492 in __db_poff () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#6  0x00007f0610577f1c in __ham_add_el () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#7  0x00007f061056f4ad in ?? () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#8  0x00007f06105fe884 in __dbc_iput () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#9  0x00007f061060ce31 in __dbc_put_pp () from /usr/lib/x86_64-linux-gnu/libdb-5.3.so
#10 0x00007f06111705a5 in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#11 0x00007f061117655f in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#12 0x00007f061117ae52 in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#13 0x00007f061118c7cb in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#14 0x00007f061118cb6f in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#15 0x00007f061118d642 in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#16 0x00007f061119d0e2 in ?? () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#17 0x00007f06111a3342 in rpmtsRun () from /usr/lib/x86_64-linux-gnu/librpm.so.3
#18 0x00007f06113dedb1 in ?? () from /usr/lib/python2.7/dist-packages/rpm/_rpmmodule.so
#19 0x00000000004dc9cb in PyEval_CallObjectWithKeywords ()
#20 0x000000000058840c in ?? ()
#21 0x0000000000499be5 in PyEval_EvalFrameEx ()
#22 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#23 0x00000000004a090c in PyEval_EvalCodeEx ()
#24 0x000000000049ab45 in PyEval_EvalFrameEx ()
#25 0x0000000000499ef2 in PyEval_EvalFrameEx ()
#26 0x00000000004a090c in PyEval_EvalCodeEx ()
#27 0x000000000049ab45 in PyEval_EvalFrameEx ()
#28 0x00000000004a090c in PyEval_EvalCodeEx ()
#29 0x0000000000499a52 in PyEval_EvalFrameEx ()
#30 0x00000000004a1634 in ?? ()
#31 0x000000000044e4a5 in PyRun_FileExFlags ()
#32 0x000000000044ec9f in PyRun_SimpleFileExFlags ()
#33 0x000000000044f904 in Py_Main ()
#34 0x00007f0612231f45 in __libc_start_main () from /lib/x86_64-linux-gnu/libc.so.6
#35 0x0000000000578c4e in _start ()
(gdb)

About this issue

  • Original URL
  • State: closed
  • Created 10 years ago
  • Comments: 19

Most upvoted comments

My persistence will have to end here for now - I may end up coming back to it once current release work eases as I would love to find out where the occasional success scenario comes from. I’m curious to see whether manually extending the file via dd if=/dev/zero of=rpmdb-file bs=1 count=0 seek=4M or similar will work.