Postings by Steve Dickson

Question about libtirpc CNAME Change.


In a recent upstream commit, the outdated, unsupported
AUTH_DES authentication is now compiled out by default.

This does effect the API, but does it really need a CNAME change
since it interface can be recompiled back in?

Changing the maintainer on a couple packages


I would like to change the maintership of nfsometer
and nfstest now that they have been re-write in
python 3.

How do I do that?



sign_and_send_pubkey: signing failed: agent refused operation


I'm getting the following error when I'm access the fedora git trees.

sign_and_send_pubkey: signing failed: agent refused operation
... at pkgs dot Permission denied (publickey).
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

Now I know I have the correct publickey (id_rsa/
because they work on a different host. and generally
when I have the wrong keys, I get the above error minus

sign_and_send_pubkey: signing failed: agent refused operation

Any idea what is happening?

Did the XML API change?


I'm getting a bunch of xml routines undefined

<a href=";volume=DEFAULT&amp;name=build.log&amp;offset=-4000" title=";volume=DEFAULT&amp;name=build.log&amp;offset=-4000">;volume=DEFAU...</a>

on my rawhide build... everything builds just fine on f29...

Fedora 30: Deprecating /etc/sysconf/nfs


A few years back there was a movement the NFS community
to consulate all nfs configuration into one file
call /etc/nfs.conf... See nfs.conf(5)

Maybe stupidly, I've maintain backwards compatibility
for that last couple Fedora releases. I think it is
time to go to the single file configuration, since
the development on Fedora 29 is winding down and
it's winding up for Fedora 30.

On fresh rawhide installs /etc/sysconfig/nfs will still
be installed but with direction to use /etc/nfs.conf

If /etc/sysconfig/nfs does exists the configuration will
*not* be overridden...

The update system has the wrong package


I'm trying to update the rpcsvc-proto package.
When I try to create a new update, the system
only gives me rpcsvc-proto-devel as a choice
not rpcsvc-proto.

Where do I open up a ticket to get this taken
care of?

Problem: package fwupdate-libs-10-6.fc28.x86_64 requires shim


I've getting the following error when I try to update from f27 to f28:

# dnf system-upgrade download --refresh --releasever=28
Before you continue ensure that your system is fully upgraded by running "dnf --refresh upgrade".

libevent-2.1.8 SONAME change.


Yesterday I updated libevent to the latest upstream release.

I mistakenly did not realized there was a SONAME change
in this update. So if your package is dependent on
libevent, you are going to have to rebuild.

My apologies for this oversight...


Package Question


Is it a problem for a package to pull from two different
upstream tar balls? Basically have

Source0: <a href="" title=""></a>
Source1: <a href="" title=""></a>

Then I would, by hand, untar Source1 into Source0 directory.

Before do the work I want to make sure I'm not
breaking violating any package rules. I did look
around and didn't see anything addressing this.

If this is kosher, are there any examples of other
packages doing this...



nfs-utils: does not compile in rawhide

Just an FYI...

nfs-utils currently does not compile in rawhide do
to this change:
<a href="" title=""></a>

We know about and should be resolved by the EOD...

Personally I think this is step in the right direction
but apologize for the inconvenience


SONAME Change: nfs-utils-2.3.1


In the latest nfs-utils upstream release the SONAME has
change on the library. The API is
was tidied up so internal functions and variable
are no longer exposed.

As I see it, this change should only effect two patches
nfs-ganesha and sssd-nfs-idmap, unless there are other
packages that are depended this library that I don't
know of.

This will only effect rawhide. I'll do the push but
wait to do the build until I talk with owners of
those packages.

What to I have to do....


What do I have to do to stop random people
from making random changes to packages I maintain?

Packaging Question


In an upcoming release the libnfsdimap library
will be rolled into the nfs-utils package.
Meaning nfs-utils will be install libnfsidmap
instead of the libnfsidmap package.

The libnfsidmap name will stay the same
so I'm hoping there will not be any problems.

What happen to UINT16_MAX


Doing koji build of nfs-utils in rawhide I'm getting
the following errors:

rpc.c: In function 'nsm_recv_getport':
rpc.c:469:13: error: 'UINT16_MAX' undeclared (first use in this function); did you mean 'UINT_MAX'?
if (port > UINT16_MAX) {

f26 and f25 build and when I clone the nfs-utils tree
on rawhide vm that build work fine as well.

Any idea what is going on?



gssproxy: server localhost not responding, timed out


Does anybody know what $subject means? I'm seeing it
with the latest rawhide while doing secure NFS testing.



nfs-utils-2.1.1 Changes Everything!


The latest nfs-utils release drastically changes how the NFS
servers are configured, for the good IMHO...

All daemon configuration now goes through /etc/nfs.conf.
See nfs.conf(5) for details.

The command line interfaces in the systemd services files
have been removed. Which means all your current configures
will break, because the variables in /etc/sysconfig/nfs are
no longer used.

Again, I think is a move in the right direction and I know
you might find this surprising 8-) but I really don't want to
break all the current server configuration.

fedup upload failure


I'm getting the following error when I'm uploading a new tarball

$ fedpkg upload libnfsidmap-0.27.tar.bz2
Could not execute upload: Can not upload a new source file with a sha512 hash, as the "/home/src/fc/libnfsidmap/sources" file contains at least one line with a md5 hash.

Please redo the whole "/home/src/fc/libnfsidmap/sources" file using: `/usr/bin/fedpkg new-sources file1 file2 ...`

when use fedpkg new-sources I get
$ fedpkg new-sources
Could not execute new_sources: Request is unauthorized.

Any ideas as to why I can not do an fedpkg upload?

Creating cores in f25


How do I get f25 to create cores, these days?

I'm getting the segfault
[55108.290610] rpc.gssd[13264]: segfault at 0 ip 000055dc90af9dde sp 00007f9fb73cb7c0 error 4 in rpc.gssd[55dc90af3000+14000]

but no core so those address are meaningless.

Everything in the kernel is set:
f25# sysctl -a | grep kernel.core
kernel.core_pattern = |/usr/libexec/abrt-hook-ccpp %s %c %p %u %g %t %P %I
kernel.core_pipe_limit = 4
kernel.core_uses_pid = 1

The abrtd daemon is up and running
f25# ps ax | grep abrtd
713 ?

Creating an new update is failing


I'm trying to create a new update and
I'm getting this error:

Builds : Unable to create update. Parent instance is not bound to a Session; lazy load operation of attribute 'release' cannot proceed

The build looks fine:
<a href="" title=""></a>

Any ideas what the problem is?



Changing the paths on binaries


Are there any rules regarding changing the
path of binaries?

rpcbind/rpcinfo has lived in /usr/sbin for
years but that is because the spec file moves
the binary from /usr/bin to /usr/sbin during
the install. Upstream defaults to /usr/bin.

Due to some systemd service file changes
I would like to just leave rpcbind/rpcinfo
in /usr/bin which means the path of those
binaries would change.



Where did res_querydomain go in f24??


This is regard to bz1372136... as the bz says

From Koji logs :
- x86_64 and armv7hl have an issue in configure : checking for res_querydomain in -lresolv... no
- i686 works : checking for res_querydomain in -lresolv... yes
This is strange, since the symbol in present in /lib/

Here was has res_querydomain or even res_nquerydomain moved to to??
Why can't I find it with a AC_CHECK_FUNC() or a AC_CHECK_LIB() macros?



STOP messing with nfs-utls... PLEASE!!!

I realized over the weekend the following
commits were make to fix the build and
they did not!! The build is just as
broken as before!!!

Now I have to go back an figure out
what has changed which takes even
more time!!!

Its not like I'm not around or unreachable
So please stop these fly by night comments
and let me do my job!

Build Error: Trigger fired by the same package is already fired???


The following build
<a href="" title=""></a>

is failing due to
error: line 256: Trigger fired by the same package is already defined in spec file: %triggerun -- nfs-utils < 1:1.3.1-4.0

Here all the trigger (that have been around forever):

triggerun -- nfs-utils < 1:1.2.4-2
/bin/systemctl enable nfs-lock.service >/dev/null 2>&1 || :
if /sbin/chkconfig --level 3 nfs ; then
/bin/systemctl enable nfs-server.service >/dev/null 2>&1 || :

%triggerin -- nfs-utils < 1:1.3.0-0.2
/bin/systemctl restart nfs-config >/dev/null 2>&1 || :


Upcoming soname change for libtirpc


Well place sources tell me in the upcoming
libtirpc release the soname will be changing
due to some API changes.

The API changes will make the Linux version
more compatible with litbirpc implementations

Obviously this change will just go in rawhide
but how do I
1) request the build rpm go be added to the build root?
2) find and rebuild all the dependencies?
3) anything else I need to do?



NetworkManager Bridges in Fedora


Has anybody had any luck with getting bridges
consistently up in running in either F19 or F20?
I know I have not...

I go into setup/networks. Add a bridge which creates
two file in /etc/sysconfig/network-scripts


The contents seem reasonable.

After I reboot, I go back in to setup/networks. The status
of the bridge is "connecting" and the Bridge slaves
are "none"???

Anybody know what has to happen to get this code working??



P.S. If there is a better list for me to ask this
question, please point me to it...

Strange Build problem....


I'm seeing following build error
<a href="" title=""></a>

But I'm not seeing this error on local f18 or f17 builds...

Note, the same version of python-matplotlib (1.2.0-5) is being used in all the builds...

How do I debug something like this??



Koji Login


For a while now I have not been able to login into
Koji... I can build packages but I can not login
to cancel any builds which is a bit annoying...

systemd automounts

I'm trying to automount /var/lib/nfs/rpc_pipefs
for the nfs-idmap.service

var-lib-nfs-rpc_pipefs.mount is:
Description=RPC Pipe File System


var-lib-nfs-rpc_pipefs.automount is:
Description=RPC Pipe File System


and the nfs-idmap.service is:
Description=Name to UID/GID mapping for NFSv4. var-lib-nfs-rpc_pipefs.automount


systemd: Is it wrong?


One of the maintainers of systemd and I have been working
together on trying to convert the NFS SysV init scripts
into systemd services. Here is the long trail...

<a href="" title=""></a>

The point is this, with fairly complicated system,
some events need to have happen, like loading modules,
before other events happen, like setting parameters of
those loaded modules.

Currently the ExecStart commands can be started and end
before the ExecStartPre even start.

Howto: Create a new package and retiring a package


Currently the nfs-utils-lib package has two libraries,
libnfsidmap and librpcsecgss. librpcsecgss is no longer
needed since it was functionally replaced by libtirpc and
now that I'm the upstream maintainer of libnfsidmap,
I would like make that its own standalone package.

So what/where are the steps I need to take to retire
nfs-utils-lib and create a new libnfsidmap package...