Rpm %post execv operations talking forever after upgrading host to fedora 31

General
1

After upgrading to Fedora 31 and running docker build I’m noticing the rpm install taking forever.

It’s stalled here:

#11 4.597 D: %post(mpfr-3.1.1-4.el7.x86_64): scriptlet start
#11 4.598 D: %post(mpfr-3.1.1-4.el7.x86_64): execv(/sbin/ldconfig) pid 11

or here

#11 4.570 D: %post(jdk1.8-2000:1.8.0_211-fcs.x86_64): scriptlet start
#11 4.570 D: %post(jdk1.8-2000:1.8.0_211-fcs.x86_64): execv(/bin/sh) pid 10

Once it starts it runs fine. This happens for both my el6 and el7 (Oracle Linux) builds. I tried checking the commands it runs and running them standalone and they run fine.

Not sure where to start looking from here? Anyone point me in the right direction?

2

Further notes on this… interesting I just installed F29 for an unrelated project and it works fine and rpm works fine. It’s only the Oracle linux rpm having a problem. Starting to wonder if it’s a command in the Dockerfile or OL specific package.

Seems OL6/7 and CentOS6 have the same problem. RPM stalls for a long time applying the %post.

3

Sorry for interrupt. but any update on this?

i’ve got same issue on this. based on centos:7.9.2009 aarch64

D: %post(libdrm-2.4.97-2.el7.aarch64): scriptlet start
D: %post(libdrm-2.4.97-2.el7.aarch64): execv(/sbin/ldconfig) pid 62

i symlink ldconfig as /usr/bin/true and no help.
strace on pid 62 gives looping of
fcntl( xxxxx) = -1 EBADF (Bad file descriptor)
but the same image works fine at another server.