Bug 21886

Summary:	Install borks because of "bad rpms"
Product:	Mageia	Reporter:	Frank Griffin <ftg>
Component:	Installer	Assignee:	Thierry Vignaud <thierry.vignaud>
Status:	RESOLVED FIXED	QA Contact:
Severity:	critical
Priority:	High	CC:	bittwister2, davidwhodgins, mageia, marja11, mhrambo3501, ngompa13, pterjan, rpmstack, thierry.vignaud, thomas.bigot, tmb, watersnowrock, zen25000
Version:	Cauldron
Target Milestone:	---
Hardware:	All
OS:	Linux
See Also:	https://bugs.mageia.org/show_bug.cgi?id=22266 https://bugs.mageia.org/show_bug.cgi?id=21971
Whiteboard:
Source RPM:	drakx-installer-stage2, rpm-4.14, perl-URPM	CVE:
Status comment:
Attachments:	Installer errors Install errors ddebug.log.xz xorg.log from an older install xorg.log from today install

Description Frank Griffin 2017-10-17 19:09:04 CEST Comment hidden (obsolete)

The current cauldron net nonfree 64-bit install aborts immediately at package selection with the errors shown in the attached screenshot.

The install is being done from a local mirror of mirror.math.princeton.edu.

Comment 1 Frank Griffin 2017-10-17 19:10:40 CEST Comment hidden (obsolete)

Created attachment 9734 [details]
Installer errors

Comment 2 Marja Van Waes 2017-10-17 23:17:06 CEST Comment hidden (obsolete)

Hi Frank,

Thanks for the report. Please reproduce the problem and then 

* switch to tty2 with Ctrl+Alt+F2

* insert a USB key

* type: 
        bug

* fetch report.bug that's written to the USB key, compress it with xz if needed

* attach report.bug(.xz) to this report

Thanks :-)

CC: (none) => marja11
Keywords: (none) => NEEDINFO

Comment 3 Frank Griffin 2017-10-18 20:43:31 CEST Comment hidden (obsolete)

Odd.  I waited for the mirror to settle several times yesterday, yet got the error.  Today when I went to reproduce it, it didn't occur.  Some metadata must have arrived later overnight which put it right.

Resolution: (none) => INVALID
Status: NEW => RESOLVED

Comment 4 Frank Griffin 2017-10-18 21:10:23 CEST

Spoke too soon.  We get through package selection and begin install, but then get a slew of "Installation failed; bad rpm" messages.

Busybox (tty2) seems totally borked.  At the top of the screen it says DrakX V17.91, but most commands, including ls, df, and bug, aren't found.

What does "bad rpm" mean ?

Status: RESOLVED => REOPENED
Resolution: INVALID => (none)

Comment 5 Frank Griffin 2017-10-18 21:11:43 CEST

Created attachment 9737 [details]
Install errors

Attachment 9734 is obsolete: 0 => 1

Frank Griffin 2017-10-18 21:18:07 CEST

Summary: Install borks due to "missing" packages required by basesystem => Install borks due to "bad rpms"

Comment 6 Marja Van Waes 2017-10-18 22:29:31 CEST Comment hidden (obsolete)

(In reply to Frank Griffin from comment #4)
> Spoke too soon.  We get through package selection and begin install, but
> then get a slew of "Installation failed; bad rpm" messages.
> 
> Busybox (tty2) seems totally borked.  At the top of the screen it says DrakX
> V17.91, but most commands, including ls, df, and bug, aren't found.

:-(

> 
> What does "bad rpm" mean ?

The last times someone hit that (see the "perl-URPM and aria2 'bad rpms' block cauldron updates" thread on dev ml), it did not mean the rpms were bad... installing them with dnf or manually with rpm did work.

Too bad there are problems in tty2 with "bug", "ls" etc. .. or do you now find a ddebug.log in the / partition you tried to install to? If so, please attach it (compress with xz if needed)

Source RPM: (none) => drakx-installer-stage2-17.91, urpmi? perl-URPM?
Assignee: bugsquad => mageiatools
Summary: Install borks due to "bad rpms" => Install borks due to "bad rpms", in tty2 commands don't work
CC: (none) => rpmstack

Comment 7 Marja Van Waes 2017-10-18 22:30:44 CEST Comment hidden (obsolete)

(In reply to Marja van Waes from comment #6)

> 
> Too bad there are problems in tty2 with "bug", "ls" etc. .. or do you now
> find a ddebug.log in the / partition you tried to install to? If so, please
> attach it (compress with xz if needed)

Location: /root/drakx/ddebug.log

Comment 8 Pascal Terjan 2017-10-18 22:41:04 CEST Comment hidden (obsolete)

It sounds like the stage2 just needs to be rebuilt to have a more recent version of rpm

CC: (none) => pterjan

Comment 9 Pascal Terjan 2017-10-19 00:18:15 CEST Comment hidden (obsolete)

Right, as Thierry pointed out, it has already been rebuilt recently so the logs would be interesting to get the version

Comment 10 Frank Griffin 2017-10-22 22:48:30 CEST

Created attachment 9747 [details]
ddebug.log.xz

Comment 11 Nicolas Lécureuil 2017-10-28 20:27:26 CEST Comment hidden (obsolete)

* ERROR: selection failed: digikam-5.5.0-2.mga6.i586 (due to conflicts with libdigikamdatabase5-5.5.0-2.mga6.i586, due to unsatisfied libdigikamdatabase.so.5.5.0, due to unsatisfied libdigikamcore.so.5.5.0, due to unsatisfied libdigikamgui.so.5.5.0)

CC: (none) => mageia

Comment 12 Frank Griffin 2017-10-28 20:41:10 CEST

Just to give the reproducible case, this is a nonfree network install with all repositories selected and all package desktops and categories (except games) selected.

Comment 13 Thierry Vignaud 2017-10-29 11:22:50 CET Comment hidden (obsolete)

This is a packaging issue, not an installer issue.

Component: Installer => RPM Packages
CC: (none) => thierry.vignaud
Source RPM: drakx-installer-stage2-17.91, urpmi? perl-URPM? => (none)

Comment 14 Marja Van Waes 2017-10-29 11:44:34 CET Comment hidden (obsolete)

(In reply to Thierry Vignaud from comment #13)
> This is a packaging issue, not an installer issue.

I fail to understand how _not_ being able to issue any normal commands in tty2 during install, isn't an installer issue :-)

However I'll reassign...

(In reply to Nicolas Lécureuil from comment #11)
> * ERROR: selection failed: digikam-5.5.0-2.mga6.i586 (due to conflicts with
> libdigikamdatabase5-5.5.0-2.mga6.i586, due to unsatisfied
> libdigikamdatabase.so.5.5.0, due to unsatisfied libdigikamcore.so.5.5.0, due
> to unsatisfied libdigikamgui.so.5.5.0)

... to KDE team, for the above non-installer problem ;-)

Keywords: NEEDINFO => (none)
Assignee: mageiatools => kde
Summary: Install borks due to "bad rpms", in tty2 commands don't work => * ERROR: selection failed: digikam-5.5.0-2.mga6.i586 (Was: Install borks due to "bad rpms", in tty2 commands don't work)
Source RPM: (none) => digikam-5.5.0-2.mga6

Comment 15 Nicolas Lécureuil 2017-10-29 12:29:49 CET

because installer is "THE INSTALLER" but here seems because it does not pass to install some rpms.

This is like when a rpm does not install, this is not rpm fault :)

Comment 16 Frank Griffin 2017-10-29 15:14:49 CET

Just as a side issue, could we please get a better error message than "bad rpm" ?

Comment 17 Frank Griffin 2017-10-29 21:38:59 CET

I'm a little confused.  

For one thing, the reproducible case will display a slew of "bad rpms" but give you an OK option.  If you take it, you'll get another slew of "bad rpms".  digikam can't be in both sets ?

Second, digikam is hardly a critical package.  Why should its failure to install derail the entire install ?  We've had several cases in the past where failure of a package may interrupt the install, but you can always continue.  Not here.

I doubt that this has anything to do with digikam.  The installer is either deciding incorrectly that many rpms are "bad", or else it's now choosing to abort if any minor error is encountered.

Comment 18 Marja Van Waes 2017-10-29 23:15:05 CET Comment hidden (obsolete)

Hi Frank,

The logs end after the "choosePackages" step, the `installPackages' step doesn't seem to start at all (or nothing got logged) ... so there are no "Installation failed: bad rpms" messages in the log file :-(

I agree the digikam errors in the "choosePackages" step are probably unrelated.  

And I think the not working commands in tty2 are also a separate issue.

I suggest keeping this report for digikam, and filing separate bug reports for the other issues when we manage to get more information. I hope to find time to do a network install next Tuesday.

Cheers,
marja

Comment 19 Frank Griffin 2017-11-01 18:31:17 CET

I'm resetting this to installer unless Thierry knows of a reason why digikam (or whatever) derails the install.  I'll open another bug for the tty stuff.

Also, why is the i586 digikam package being selected for a 64-bit install ?

Resetting assignee since doesn't seem to have anything to do with KDE.

Component: RPM Packages => Installer
Assignee: kde => bugsquad
Summary: * ERROR: selection failed: digikam-5.5.0-2.mga6.i586 (Was: Install borks due to "bad rpms", in tty2 commands don't work) => * Install aborts; ERROR: selection failed: digikam-5.5.0-2.mga6.i586

Comment 20 Marja Van Waes 2017-11-01 19:14:45 CET Comment hidden (obsolete)

Originally, the 64bit digikam package was selected.

* selecting digikam-5.5.0-2.mga6.x86_64

some dependencies were selected and then it went wrong here:

* selecting lib64digikamcore5-5.5.0-2.mga6.x86_64
* requiring libKF5Sane.so.5()(64bit),liblqr-1.so.0()(64bit),libmarblewidget-qt5.so.26()(64bit),libopencv_objdetect.so.2.4()(64bit) for lib64digikamcore5-5.5.0-2.mga6.x86_64
* unselecting lib64digikamcore5-5.5.0-2.mga6.x86_64
* unselecting lib64digikamdatabase5-5.5.0-2.mga6.x86_64
* unselecting digikam-5.5.0-2.mga6.x86_64
* unselecting lib64digikamgui5-5.5.0-2.mga6.x86_64
* adding a reason to already rejected package lib64digikamcore5-5.5.0-2.mga6.x86_64: unsatisfied libmarblewidget-qt5.so.26()(64bit)
* no packages match libdigikamcore.so.5.5.0()(64bit) (it is either in skip.list or already rejected)
* adding a reason to already rejected package lib64digikamgui5-5.5.0-2.mga6.x86_64: unsatisfied libdigikamcore.so.5.5.0()(64bit)
* no packages match libdigikamdatabase.so.5.5.0()(64bit) (it is either in skip.list or already rejected)
* adding a reason to already rejected package digikam-5.5.0-2.mga6.x86_64: unsatisfied libdigikamdatabase.so.5.5.0()(64bit)
* no packages match libdigikamcore.so.5.5.0()(64bit) (it is either in skip.list or already rejected)
* adding a reason to already rejected package digikam-5.5.0-2.mga6.x86_64: unsatisfied libdigikamcore.so.5.5.0()(64bit)

So, iinm there wouldn't have been a problem with digikam if libmarblewidget-qt5.so.26()(64bit) had been satisfied.

However, I still don't believe those errors are related to install borking after the package selection step finished.

Comment 21 Frank Griffin 2017-11-04 15:23:03 CET Comment hidden (obsolete)

Same behavior with the Nov 3 stage2.

Comment 22 Frank Griffin 2017-11-18 14:45:01 CET Comment hidden (obsolete)

Same behavior with the Nov 11 ISOs.  However, I notice that mdkinst.sqfs still has a timestamp of Oct 18.  Is this the problem ?

Comment 23 Frank Griffin 2017-11-30 20:25:37 CET

Same behavior with the Nov 30 stage2.

Comment 24 Marja Van Waes 2017-12-02 08:30:24 CET

(In reply to Frank Griffin from comment #23)
> Same behavior with the Nov 30 stage2.

What exactly is "same behaviour"

I'm reverting to the previous summary of this bug report, because rereading this report, it seems I missed or misread neoclust's last comment.

Is all of 'Install borks due to "bad rpms", in tty2 commands don't work' still correct?

Summary: * Install aborts; ERROR: selection failed: digikam-5.5.0-2.mga6.i586 => Install borks due to "bad rpms", in tty2 commands don't work
Assignee: bugsquad => mageiatools

Marja Van Waes 2017-12-02 08:30:56 CET

Source RPM: digikam-5.5.0-2.mga6 => drakx-installer-stage2

Comment 25 Frank Griffin 2017-12-02 17:17:33 CET

(In reply to Marja van Waes from comment #24)
> (In reply to Frank Griffin from comment #23)
> > Same behavior with the Nov 30 stage2.
> 
> What exactly is "same behaviour"

The "bad rpms" error.

> 
> Is all of 'Install borks due to "bad rpms", in tty2 commands don't work'
> still correct?

Just the first part; I opened https://bugs.mageia.org/show_bug.cgi?id=21971 for the tty2 commands.

Frank Griffin 2017-12-02 20:48:26 CET

Summary: Install borks due to "bad rpms", in tty2 commands don't work => Install borks because thousands of packages required by grub2 can't install

Comment 26 Frank Griffin 2017-12-02 20:55:31 CET Comment hidden (obsolete)

The mdkinst.sqfs built for bug#21971 has fixed this, but surfaced another error.  Package Selection aborts immediately with a popup error "Some packages required by grub2 cannot be installed and then proceeds to list practically every package in the repository (it seems).

tty2 still doesn't work.  I'll see if I can extract ddebug.log.

Comment 27 Thomas Backlund 2017-12-02 21:17:20 CET Comment hidden (obsolete)

(In reply to Frank Griffin from comment #26)
> The mdkinst.sqfs built for bug#21971 has fixed this, but surfaced another
> error.  Package Selection aborts immediately with a popup error "Some
> packages required by grub2 cannot be installed and then proceeds to list
> practically every package in the repository (it seems).
> 


probably a fallout of missing libreadline6.so

I've restored the missing rpms to the mirror so we can rebuild the rest of the packages depending on old readline

CC: (none) => tmb

Comment 28 Frank Griffin 2017-12-05 20:03:55 CET

With today's cauldron, we're back to "bad rpms"...

Summary: Install borks because thousands of packages required by grub2 can't install => Install borks because of "bad rpms"

Comment 29 Barry Jackson 2017-12-13 19:28:55 CET

I just attempted a vbox install of cauldron using latest Mageia-Cauldron-netinstall-x86_64.iso dated 6th December and hit the same bad rpms issue.
Can get into a terminal but no commands work:
bash-4.4# bug
bash: bug: command not found

same for ls rpm df however echo works :)

OK-ing the GUI error message over and over and stopping on the final failure screen I found a message on tty3 after a list of the full urls of all the packages:

Try to continue anyway?
* opening rpmdb (root=/mnt, write=1))
* workaround bug in rpmlib by removing /mnt/var/lib/rpm/__db*
* 22 installation transactions failed
*

CC: (none) => zen25000

Comment 30 Frank Griffin 2017-12-15 18:39:39 CET Comment hidden (obsolete)

Same problem with Dec 14 uploads.  mdkinst and rescue are still at Dec 2.

Comment 31 Frank Griffin 2017-12-18 17:09:33 CET

Same problem still with Dec 17 stage2 and mdkinst.

Marja Van Waes 2017-12-25 21:08:10 CET

See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=22266

Comment 32 Dave Hodgins 2017-12-26 09:07:52 CET

Looks like something is wrong with the creation of the squashfs file.

Using http://mirror.math.princeton.edu/pub/mageia/distrib/cauldron/x86_64/install/stage2/mdkinst.sqfs

[root@x3 tmp]# unsquashfs -s mdkinst.sqfs 
Found a valid SQUASHFS 4:0 superblock on mdkinst.sqfs.
Creation or last append time Sun Dec 17 03:21:24 2017
Filesystem size 70753.46 Kbytes (69.10 Mbytes)
Compression xz
Block size 131072
Filesystem is exportable via NFS
Inodes are compressed
Data is compressed
Fragments are compressed
Always-use-fragments option is not specified
Xattrs are compressed
Duplicates are removed
Number of fragments 483
Number of inodes 4414
Number of ids 1
[root@x3 tmp]# unsquashfs -l mdkinst.sqfs 
Parallel unsquashfs: Using 4 processors
Read on filesystem failed because EOF
read_uids_guids: failed to read id index table
FATAL ERROR:failed to uid/gid table

So the squashfs file is not fully readable.
The same command using the m6 version for the mdkinst.sqfs file does list the contents of the file.

CC: (none) => davidwhodgins

Comment 33 Marja Van Waes 2017-12-26 09:21:49 CET

w00t!

Thanks a lot, Dave!

Summary: Install borks because of "bad rpms" => Install borks because of corrupted mdkinst.sqfs

Marja Van Waes 2017-12-26 09:30:18 CET

See Also: (none) => https://bugs.mageia.org/show_bug.cgi?id=21971

Marja Van Waes 2017-12-26 09:37:41 CET

Priority: Normal => High
Severity: major => critical

Comment 34 Frank Griffin 2018-01-04 00:42:33 CET

OK, so the mdkinst.sqfs is corrupted, probably truncated due to some space constraint.  Could we please fix this so that cauldron installs are once again possible ?

Comment 35 Pascal Terjan 2018-01-04 05:31:22 CET

I don't think there is a problem on creation of the squashfs, it seems valid here (tested using 
unsquashfs -l /usr/lib64/drakx-installer-stage2/install/stage2/mdkinst.sqfs)

No problem either when downloading it from //mirror.math.princeton.edu/pub/mageia/distrib/cauldron/x86_64/install/stage2/mdkinst.sqfs

(In reply to Barry Jackson from comment #29)
> Can get into a terminal but no commands work:
> bash-4.4# bug
> bash: bug: command not found
> 
> same for ls rpm df however echo works :)

This part should be fixed by http://gitweb.mageia.org/software/drakx/commit/?id=d24876d991bbcfb83fd523896dba85e2df1fce46 (but I didn't see this comment before fixing it so didn't link the commit to this bug)

Comment 36 Thomas Backlund 2018-01-04 14:51:49 CET

I've pushed a drakx-installer-stage2-17.92-1.mga7 with the fixes by pterjan to the mirrors...

Feel free to test when cauldron/x86_64/install/stage2/VERSION states 17.92 on the mirror you use

Comment 37 Frank Griffin 2018-01-04 17:25:03 CET

I've retested with Pascal's fix and I confirm that the tty2 busybox problem in bug#21971 is fixed, but the bad rpms issue persists.

However, I did "unsquashfs -l mdkinst.sqfs" and got correct output, so Pascal is correct - it's not that.

Comment 38 Frank Griffin 2018-01-05 17:45:16 CET

Still failing with today's stage2 upload.

Frank Griffin 2018-01-05 17:46:00 CET

Summary: Install borks because of corrupted mdkinst.sqfs => Install borks because of "bad rpms"

Comment 39 Donald 2018-01-06 03:40:26 CET

Failed for me today, was also unable to update Cauldron due to bad rpms, dnf also failed - unable to sync caches.

The update was to urpmi and rpm itself, could the error be in one of those?

CC: (none) => watersnowrock

Thomas Bigot 2018-01-08 13:42:42 CET

CC: (none) => thomas.bigot

Comment 40 Donald 2018-01-10 14:23:02 CET

I did a full local resync of a mirror that was giving bad rpms, the install and subsequent updates have been trouble free since then. 

Not sure how relevant this is, but might help trace the issue down.

Comment 41 Frank Griffin 2018-01-15 02:03:04 CET

(In reply to Donald from comment #40)
> I did a full local resync of a mirror that was giving bad rpms, the install
> and subsequent updates have been trouble free since then. 
> 
> Not sure how relevant this is, but might help trace the issue down.

Please elaborate on "full local resync".  I sync my local mirror from the tier 1 princeton mirror using rsync.  Are you saying that rsync -avHL creates files which are not identical to the source mirror ?

@Barry
Were you using a netinstall from a tier 1 mirror or from a local copy of one ?

I'm really hesitant to wipe my local mirror and force a complete resync of cauldron without some better evidence.

Comment 42 Donald 2018-01-15 03:25:17 CET

(In reply to Frank Griffin from comment #41)
> (In reply to Donald from comment #40)
> > I did a full local resync of a mirror that was giving bad rpms, the install
> > and subsequent updates have been trouble free since then. 
> > 
> > Not sure how relevant this is, but might help trace the issue down.
> 
> Please elaborate on "full local resync".  I sync my local mirror from the
> tier 1 princeton mirror using rsync.  Are you saying that rsync -avHL
> creates files which are not identical to the source mirror ?

I sync with ftp-stud.hs-esslingen.de, and I seriously doubt that the rsync gives different files - more than a little weight against that. More likely, there was either an update between the resync or some installer quirk.

I haven't tried reinstalling over the network directly since then, however, updates/installing packages using urpmi over the network have not been an issue.

DNF remains broken with the above caching issue.

What netinstall image were you using? I had success with the latest, dated January 10th.

> 
> @Barry
> Were you using a netinstall from a tier 1 mirror or from a local copy of one
> ?
> 
> I'm really hesitant to wipe my local mirror and force a complete resync of
> cauldron without some better evidence.

Comment 43 Donald 2018-01-15 03:27:39 CET

(In reply to Donald from comment #42)
> (In reply to Frank Griffin from comment #41)
> > (In reply to Donald from comment #40)
> > > I did a full local resync of a mirror that was giving bad rpms, the install
> > > and subsequent updates have been trouble free since then. 
> > > 
> > > Not sure how relevant this is, but might help trace the issue down.
> > 
> > Please elaborate on "full local resync".  I sync my local mirror from the
> > tier 1 princeton mirror using rsync.  Are you saying that rsync -avHL
> > creates files which are not identical to the source mirror ?
> 
> I sync with ftp-stud.hs-esslingen.de, and I seriously doubt that the rsync
> gives different files - more than a little weight against that. More likely,
> there was either an update between the resync or some installer quirk.

Forgot to add, this was a completely fresh rsync, so maybe there is some file leftover that is causing a bug somewhere, but I can't see how this could realistically be the case!

> 
> I haven't tried reinstalling over the network directly since then, however,
> updates/installing packages using urpmi over the network have not been an
> issue.
> 
> DNF remains broken with the above caching issue.
> 
> What netinstall image were you using? I had success with the latest, dated
> January 10th.
> 
> > 
> > @Barry
> > Were you using a netinstall from a tier 1 mirror or from a local copy of one
> > ?
> > 
> > I'm really hesitant to wipe my local mirror and force a complete resync of
> > cauldron without some better evidence.

Comment 44 Marja Van Waes 2018-01-15 08:51:57 CET

(In reply to Donald from comment #39)
> Failed for me today, was also unable to update Cauldron due to bad rpms, dnf
> also failed - unable to sync caches.
> 
> The update was to urpmi and rpm itself, could the error be in one of those?


(In reply to Donald from comment #42)

> 
> DNF remains broken with the above caching issue.
> 

Does the following, as root, help?:

    dnf clean all

CC: (none) => ngompa13

Comment 45 Mike Rambo 2018-01-17 17:05:21 CET

I've been running into the 'bad rpms' problem for several days now trying to install a new cauldron (Vbox). cauldron/x86_64/install/stage2/VERSION states 17.92 on the princeton mirror I use. I've also copied the mirror locally and have the same failure. I guess that makes me a good test subject if anyone has some ideas to try.

CC: (none) => mrambo

Comment 46 Pascal Terjan 2018-01-19 21:15:29 CET

Problem happens immediately in the first transaction:

* retrieving   icu59-data-59.1-5.mga7.noarch.rpm lib64lzma5-5.2.3-2.mga7.x86_64.rpm lib64ffi6-3.2.1-5.mga6.x86_64.rpm bash-4.4-12.1.mga7.x86_64.rpm lib64popt0-1.16-13.mga6.x86_64.rpm lib64pcre1-8.41-3.mga7.x86_64.rpm lib64blkid1-2.30.1-2.mga7.x86_64.rpm telepathy-filesystem-0.0.2-6.mga6.noarch.rpm shared-mime-info-1.8-2.mga7.x86_64.rpm lib64glib2.0_0-2.54.3-1.mga7.x86_64.rpm lib64gdbm4-1.13-1.mga7.x86_64.rpm lib64systemd0-236-4.mga7.x86_64.rpm lib64gio2.0_0-2.54.3-1.mga7.x86_64.rpm lib64gpg-error0-1.27-1.mga7.x86_64.rpm lib64uuid1-2.30.1-2.mga7.x86_64.rpm lib64zlib1-1.2.11-4.mga6.x86_64.rpm lib64lz4_1-1.8.0-2.mga7.x86_64.rpm lib64ncurses6-6.0-20171223.2.mga7.x86_64.rpm lib64gmp10-6.1.2-2.mga7.x86_64.rpm gdisk-1.0.1-1.mga6.x86_64.rpm perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm dash-static-0.5.9-2.mga7.x86_64.rpm libstdc++6-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm xz-5.2.3-2.mga7.x86_64.rpm lib64db5.3-5.3.28-14.mga7.x86_64.rpm lib64xml2_2-2.9.7-2.mga7.x86_64.rpm setup-2.7.24-1.mga6.noarch.rpm glibc-2.22-26.mga7.x86_64.rpm perl-base-5.26.1-2.mga7.x86_64.rpm pkgconfig-0.29.2-1.mga7.x86_64.rpm lib64cap2-2.25-1.mga6.x86_64.rpm lib64gcrypt20-1.8.1-1.mga7.x86_64.rpm filesystem-2.1.9-28.mga6.x86_64.rpm perl-5.26.1-2.mga7.x86_64.rpm perl-Time-HiRes-1.975.0-1.mga7.x86_64.rpm lib64mount1-2.30.1-2.mga7.x86_64.rpm run-parts-4.4-7.mga6.x86_64.rpm lib64acl1-2.2.52-7.mga6.x86_64.rpm grep-3.1-1.mga7.x86_64.rpm lib64icu59-59.1-5.mga7.x86_64.rpm coreutils-8.28-1.mga7.x86_64.rpm popt-data-1.16-13.mga6.noarch.rpm libgpg-error-common-1.27-1.mga7.noarch.rpm lib64sqlite3_0-3.21.0-2.mga7.x86_64.rpm lib64signon-glib1-1.7-14.mga6.x86_64.rpm lib64attr1-2.4.48-1.mga7.x86_64.rpm libgcc1-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm lib64accounts-glib0-1.23-1.mga6.x86_64.rpm
* '/usr/bin/curl' '-q' '--location-trusted' '-R' '-f' '--disable-epsv' '--connect-timeout' '60' '--anyauth' '--stderr' '-' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/icu59-data-59.1-5.mga7.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64lzma5-5.2.3-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64ffi6-3.2.1-5.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/bash-4.4-12.1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64popt0-1.16-13.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64pcre1-8.41-3.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64blkid1-2.30.1-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/telepathy-filesystem-0.0.2-6.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/shared-mime-info-1.8-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64glib2.0_0-2.54.3-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gdbm4-1.13-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64systemd0-236-4.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gio2.0_0-2.54.3-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gpg-error0-1.27-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64uuid1-2.30.1-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64zlib1-1.2.11-4.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64lz4_1-1.8.0-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64ncurses6-6.0-20171223.2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gmp10-6.1.2-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/gdisk-1.0.1-1.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/dash-static-0.5.9-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/libstdc++6-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/xz-5.2.3-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64db5.3-5.3.28-14.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64xml2_2-2.9.7-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/setup-2.7.24-1.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/glibc-2.22-26.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-base-5.26.1-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/pkgconfig-0.29.2-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64cap2-2.25-1.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gcrypt20-1.8.1-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/filesystem-2.1.9-28.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-5.26.1-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-Time-HiRes-1.975.0-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64mount1-2.30.1-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/run-parts-4.4-7.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64acl1-2.2.52-7.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/grep-3.1-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64icu59-59.1-5.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/coreutils-8.28-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/popt-data-1.16-13.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/libgpg-error-common-1.27-1.mga7.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64sqlite3_0-3.21.0-2.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64signon-glib1-1.7-14.mga6.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64attr1-2.4.48-1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/libgcc1-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64accounts-glib0-1.23-1.mga6.x86_64.rpm'
* retrieving   libdb5.3-5.3.28-14.mga7.i586.rpm libzlib1-1.2.11-4.mga6.i586.rpm icu59-data-59.1-5.mga7.noarch.rpm setup-2.7.24-1.mga6.noarch.rpm telepathy-filesystem-0.0.2-6.mga6.noarch.rpm popt-data-1.16-13.mga6.noarch.rpm perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm libgpg-error-common-1.27-1.mga7.noarch.rpm
* '/usr/bin/curl' '-q' '--location-trusted' '-R' '-f' '--disable-epsv' '--connect-timeout' '60' '--anyauth' '--stderr' '-' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libdb5.3-5.3.28-14.mga7.i586.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libzlib1-1.2.11-4.mga6.i586.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/icu59-data-59.1-5.mga7.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/setup-2.7.24-1.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/telepathy-filesystem-0.0.2-6.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/popt-data-1.16-13.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm' '-O' 'http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libgpg-error-common-1.27-1.mga7.noarch.rpm'
* Installation failed, bad rpms:
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/libstdc++6-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/dash-static-0.5.9-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/gdisk-1.0.1-1.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gmp10-6.1.2-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64ncurses6-6.0-20171223.2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64lz4_1-1.8.0-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64zlib1-1.2.11-4.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gpg-error0-1.27-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64uuid1-2.30.1-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gio2.0_0-2.54.3-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64glib2.0_0-2.54.3-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gdbm4-1.13-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64systemd0-236-4.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/shared-mime-info-1.8-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64blkid1-2.30.1-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/telepathy-filesystem-0.0.2-6.mga6.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64pcre1-8.41-3.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64popt0-1.16-13.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/bash-4.4-12.1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64ffi6-3.2.1-5.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/icu59-data-59.1-5.mga7.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64lzma5-5.2.3-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64attr1-2.4.48-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/libgcc1-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64accounts-glib0-1.23-1.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64sqlite3_0-3.21.0-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64signon-glib1-1.7-14.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/popt-data-1.16-13.mga6.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libgpg-error-common-1.27-1.mga7.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/coreutils-8.28-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64icu59-59.1-5.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64acl1-2.2.52-7.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/grep-3.1-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/run-parts-4.4-7.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-Time-HiRes-1.975.0-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64mount1-2.30.1-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-5.26.1-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/filesystem-2.1.9-28.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libdb5.3-5.3.28-14.mga7.i586.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64cap2-2.25-1.mga6.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64gcrypt20-1.8.1-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/pkgconfig-0.29.2-1.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/setup-2.7.24-1.mga6.noarch.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/glibc-2.22-26.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/perl-base-5.26.1-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/xz-5.2.3-2.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/i586/media/core/release/libzlib1-1.2.11-4.mga6.i586.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64db5.3-5.3.28-14.mga7.x86_64.rpm
    http://ftp.free.fr/mirrors/mageia.org/distrib/cauldron/x86_64/media/core/release/lib64xml2_2-2.9.7-2.mga7.x86_64.rpm

Try to continue anyway?

Comment 47 Pascal Terjan 2018-01-19 21:19:52 CET

And from install.log, not much more:

retrieving rpm files from medium "Core Release"...
retrieved   icu59-data-59.1-5.mga7.noarch.rpm lib64lzma5-5.2.3-2.mga7.x86_64.rpm lib64ffi6-3.2.1-5.mga6.x86_64.rpm bash-4.4-12.1.mga7.x86_64.rpm lib64popt0-1.16-13.mga6.x86_64.rpm lib64pcre1-8.41-3.mga7.x86_64.rpm lib64blkid1-2.30.1-2.mga7.x86_64.rpm telepathy-filesystem-0.0.2-6.mga6.noarch.rpm shared-mime-info-1.8-2.mga7.x86_64.rpm lib64glib2.0_0-2.54.3-1.mga7.x86_64.rpm lib64gdbm4-1.13-1.mga7.x86_64.rpm lib64systemd0-236-4.mga7.x86_64.rpm lib64gio2.0_0-2.54.3-1.mga7.x86_64.rpm lib64gpg-error0-1.27-1.mga7.x86_64.rpm lib64uuid1-2.30.1-2.mga7.x86_64.rpm lib64zlib1-1.2.11-4.mga6.x86_64.rpm lib64lz4_1-1.8.0-2.mga7.x86_64.rpm lib64ncurses6-6.0-20171223.2.mga7.x86_64.rpm lib64gmp10-6.1.2-2.mga7.x86_64.rpm gdisk-1.0.1-1.mga6.x86_64.rpm perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm dash-static-0.5.9-2.mga7.x86_64.rpm libstdc++6-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm xz-5.2.3-2.mga7.x86_64.rpm lib64db5.3-5.3.28-14.mga7.x86_64.rpm lib64xml2_2-2.9.7-2.mga7.x86_64.rpm setup-2.7.24-1.mga6.noarch.rpm glibc-2.22-26.mga7.x86_64.rpm perl-base-5.26.1-2.mga7.x86_64.rpm pkgconfig-0.29.2-1.mga7.x86_64.rpm lib64cap2-2.25-1.mga6.x86_64.rpm lib64gcrypt20-1.8.1-1.mga7.x86_64.rpm filesystem-2.1.9-28.mga6.x86_64.rpm perl-5.26.1-2.mga7.x86_64.rpm perl-Time-HiRes-1.975.0-1.mga7.x86_64.rpm lib64mount1-2.30.1-2.mga7.x86_64.rpm run-parts-4.4-7.mga6.x86_64.rpm lib64acl1-2.2.52-7.mga6.x86_64.rpm grep-3.1-1.mga7.x86_64.rpm lib64icu59-59.1-5.mga7.x86_64.rpm coreutils-8.28-1.mga7.x86_64.rpm popt-data-1.16-13.mga6.noarch.rpm libgpg-error-common-1.27-1.mga7.noarch.rpm lib64sqlite3_0-3.21.0-2.mga7.x86_64.rpm lib64signon-glib1-1.7-14.mga6.x86_64.rpm lib64attr1-2.4.48-1.mga7.x86_64.rpm libgcc1-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm lib64accounts-glib0-1.23-1.mga6.x86_64.rpm
...retrieving done
retrieving rpm files from medium "Core 32bit Release"...
retrieved   libdb5.3-5.3.28-14.mga7.i586.rpm libzlib1-1.2.11-4.mga6.i586.rpm icu59-data-59.1-5.mga7.noarch.rpm setup-2.7.24-1.mga6.noarch.rpm telepathy-filesystem-0.0.2-6.mga6.noarch.rpm popt-data-1.16-13.mga6.noarch.rpm perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm libgpg-error-common-1.27-1.mga7.noarch.rpm
...retrieving done

I wonder why it is installing two i586 libraries (libdb5.3 and libzlib1) but I see no error enywhere

Comment 48 Pascal Terjan 2018-01-19 21:29:06 CET

The ton of i586 stuff is coming from:

* selecting digikam-5.5.0-2.mga6.x86_64
* requiring digikam-core[== 1:5.5.0-2.mga6],kipi-common,libdigikamcore.so.5.5.0()(64bit),libdigikamdatabase.so.5.5.0()(64bit),libdigikamgui.so.5.5.0()(64bit),libkface,mysql-common,mysql-core,sqlite3-tools for digikam-5.5.0-2.mga6.x86_64
[...]
* chosen lib64digikamdatabase5-5.5.0-2.mga6.x86_64 for libdigikamdatabase.so.5.5.0()(64bit)
* selecting lib64digikamdatabase5-5.5.0-2.mga6.x86_64
* requiring libdigikamcore.so.5.5.0()(64bit) for lib64digikamdatabase5-5.5.0-2.mga6.x86_64
* chosen lib64digikamcore5-5.5.0-2.mga6.x86_64 for libdigikamcore.so.5.5.0()(64bit)
* selecting lib64digikamcore5-5.5.0-2.mga6.x86_64
[...]
* no packages match libmarblewidget-qt5.so.26()(64bit) (it is either in skip.list or already rejected)
* unselecting lib64digikamcore5-5.5.0-2.mga6.x86_64
* unselecting digikam-5.5.0-2.mga6.x86_64
* unselecting lib64digikamdatabase5-5.5.0-2.mga6.x86_64
* adding a reason to already rejected package lib64digikamcore5-5.5.0-2.mga6.x86_64: unsatisfied libmarblewidget-qt5.so.26()(64bit)
[...]
* packageCallbackChoices: default choice ('libdigikamcore5') from libdigikamcore5,libdigikamgui5 for digikam-core[== 1:5.5.0-2.mga6]
* replacing digikam-core[== 1:5.5.0-2.mga6] with libdigikamcore5
* selecting libdigikamcore5-5.5.0-2.mga6.i586

That doesn't explain why rpm is upset but that explain the i586 packages coming in

Comment 49 Frank Griffin 2018-01-19 21:48:41 CET

Does the "bad rpm" message come from our source or from rpm ?  If from ours, it would be nice to know exactly what this means.  In fact, regardless of where it's from, it would be nice to know what it means.

Comment 50 Pascal Terjan 2018-01-19 23:52:46 CET

It comes from perl-URPM, it means parse_rpm failed.

My guess would be that some new file is missing from the installer and librpm fails to load the rpm files but no idea what so far.

Comment 51 Pascal Terjan 2018-01-20 23:29:31 CET

I could reproduce with drakx-in-chroot but if I stop it before it delete the "bad rpms", the perl-URPM from stage2 is happy with the files so I have no idea why it fails:

[root@mageia /]# perl -MURPM -e 'my $u=URPM->new(); print $u->parse_rpm("/mnt/var/cache/urpmi/partial/adwaita-icon-theme-3.26.1-1.mga7.noarch.rpm"); print "\n";'
00

Comment 52 Pascal Terjan 2018-01-20 23:30:26 CET

Also, it works fine with a mirror on local disk but not with ftp or http mirror.

Comment 53 Donald 2018-01-20 23:33:34 CET

(In reply to Pascal Terjan from comment #52)
> Also, it works fine with a mirror on local disk but not with ftp or http
> mirror.

That was exactly what I observed, both for actually installing, and for updates.

Oddly, enabling a ftp/http 32bit set of sources on a 64bit system seems to not lead to any impact on the packages installation. Note that the 32bit media were only used to install steam, so perhaps the issue is with a more core package.

Comment 54 Thomas Backlund 2018-01-21 00:15:34 CET

(In reply to Pascal Terjan from comment #52)
> Also, it works fine with a mirror on local disk but not with ftp or http
> mirror.

Hm, does downloader bits/libs return other code than before then so we flag it as "bad rpms" ?

Comment 55 Frank Griffin 2018-01-21 00:19:09 CET

Ahh, I think we're getting closer.  I use a local mirror sync'd from princeton, but I do the install using http from this local mirror because of several ancient bugs with 64-bit installs using HDD and NFS which have never been addressed (or if they have, the associated bugs were never updated).

If NFS 64-bit now works, I'll retest using that.

If this turns out to be the case, we really ought to consolidate the code between these different access methods to use common code.

Comment 56 Frank Griffin 2018-01-21 00:25:30 CET

The older bug about 64-bit NFS installs was bug#2577.

Comment 57 Pascal Terjan 2018-01-21 00:36:22 CET

According to the logs the retrieval succeeds, and install.log says:

retrieving rpm files from medium "Core Release"...
retrieved   lib64lz4_1-1.8.0-2.mga7.x86_64.rpm lib64ncursesw6-6.0-20171223.2.mga7.x86_64.rpm shared-mime-info-1.8-2.mga7.x86_64.rpm lib64openssl1.1-1.1.0g-1.mga7.x86_64.rpm lib64sqlite3_0-3.21.0-2.mga7.x86_64.rpm lib64xml2_2-2.9.7-2.mga7.x86_64.rpm xz-5.2.3-2.mga7.x86_64.rpm lib64icu59-59.1-5.mga7.x86_64.rpm filesystem-2.1.9-28.mga6.x86_64.rpm lib64glib2.0_0-2.54.3-1.mga7.x86_64.rpm lib64expat1-2.2.5-1.mga7.x86_64.rpm lib64popt0-1.16-13.mga6.x86_64.rpm lib64lzma5-5.2.3-2.mga7.x86_64.rpm libstdc++6-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm lib64db5.3-5.3.28-14.mga7.x86_64.rpm lib64attr1-2.4.48-1.mga7.x86_64.rpm perl-5.26.1-2.mga7.x86_64.rpm gdisk-1.0.1-1.mga6.x86_64.rpm lib64blkid1-2.30.1-2.mga7.x86_64.rpm lib64bz2_1-1.0.6-11.mga7.x86_64.rpm lib64systemd0-236-4.mga7.x86_64.rpm setup-2.7.24-1.mga6.noarch.rpm bash-4.4-12.1.mga7.x86_64.rpm info-install-6.1-1.mga6.x86_64.rpm lib64uuid1-2.30.1-2.mga7.x86_64.rpm lib64pcre1-8.41-3.mga7.x86_64.rpm lib64gcrypt20-1.8.1-1.mga7.x86_64.rpm perl-Time-HiRes-1.975.0-1.mga7.x86_64.rpm coreutils-8.28-1.mga7.x86_64.rpm perl-Archive-Cpio-0.100.0-1.mga6.noarch.rpm lib64gpg-error0-1.27-1.mga7.x86_64.rpm lib64cap2-2.25-1.mga6.x86_64.rpm pkgconfig-0.29.2-1.mga7.x86_64.rpm lib64ffi6-3.2.1-5.mga6.x86_64.rpm libgpg-error-common-1.27-1.mga7.noarch.rpm lib64mount1-2.30.1-2.mga7.x86_64.rpm lib64gio2.0_0-2.54.3-1.mga7.x86_64.rpm lib64acl1-2.2.52-7.mga6.x86_64.rpm lib64gmp10-6.1.2-2.mga7.x86_64.rpm lib64gdbm4-1.13-1.mga7.x86_64.rpm run-parts-4.4-7.mga6.x86_64.rpm crypto-policies-20170606-2.mga7.noarch.rpm lib64zlib1-1.2.11-4.mga6.x86_64.rpm glibc-2.22-27.mga7.x86_64.rpm lib64ncurses6-6.0-20171223.2.mga7.x86_64.rpm perl-base-5.26.1-2.mga7.x86_64.rpm sed-4.4-1.mga7.x86_64.rpm libgcc1-7.3.0-0.rc.20180117.1.mga7.x86_64.rpm icu59-data-59.1-5.mga7.noarch.rpm grep-3.1-1.mga7.x86_64.rpm
...retrieving done
starting installing packages
created transaction for installing on /mnt (remove=0, install=0, upgrade=0)

Comment 58 Pascal Terjan 2018-01-21 00:39:23 CET

I am wondering if it could be a bug like searching the files in urpmi cache outside the chroot or something like that, I'll try to strace things.

Comment 59 Pascal Terjan 2018-01-21 00:45:11 CET

Hmm no it seems to open the file just fine...

[pid 13996] stat("/mnt/var/cache/urpmi/partial/lib64attr1-2.4.48-1.mga7.x86_64.rpm", {st_mode=S_IFREG|0644, st_size=15768, ...}) = 0
[pid 13996] open("/mnt/var/cache/urpmi/partial/lib64attr1-2.4.48-1.mga7.x86_64.rpm", O_RDONLY) = 7
[pid 13996] fcntl(7, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] getgid()                    = 0
[pid 13996] getuid()                    = 0
[pid 13996] stat("/var", {st_mode=S_IFDIR|0755, st_size=140, ...}) = 0
[pid 13996] stat("/var/lib", {st_mode=S_IFDIR|0755, st_size=80, ...}) = 0
[pid 13996] stat("/var/lib/rpm", {st_mode=S_IFDIR|0755, st_size=500, ...}) = 0
[pid 13996] rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
[pid 13996] rt_sigaction(SIGINT, NULL, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGINT, {sa_handler=0x7f22a4fff5c0, sa_mask=[], sa_flags=SA_RESTORER|SA_SIGINFO, sa_restorer=0x7f22a8ffea70}, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGQUIT, NULL, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGQUIT, {sa_handler=0x7f22a4fff5c0, sa_mask=[], sa_flags=SA_RESTORER|SA_SIGINFO, sa_restorer=0x7f22a8ffea70}, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGHUP, NULL, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGHUP, {sa_handler=0x7f22a4fff5c0, sa_mask=[], sa_flags=SA_RESTORER|SA_SIGINFO, sa_restorer=0x7f22a8ffea70}, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGTERM, NULL, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGTERM, {sa_handler=0x7f22a4fff5c0, sa_mask=[], sa_flags=SA_RESTORER|SA_SIGINFO, sa_restorer=0x7f22a8ffea70}, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, 8) = 0
[pid 13996] rt_sigaction(SIGPIPE, NULL, {sa_handler=SIG_IGN, sa_mask=[PIPE], sa_flags=SA_RESTORER|SA_RESTART, sa_restorer=0x7f22a80fe7e0}, 8) = 0
[pid 13996] rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
[pid 13996] access("/var/lib/rpm/Packages", F_OK) = 0
[pid 13996] umask(022)                  = 022
[pid 13996] open("/var/lib/rpm/.dbenv.lock", O_RDWR|O_CREAT, 0644) = 8
[pid 13996] umask(022)                  = 022
[pid 13996] fcntl(8, F_SETLKW, {l_type=F_WRLCK, l_whence=SEEK_SET, l_start=0, l_len=0}) = 0
[pid 13996] open("/var/lib/rpm/DB_CONFIG", O_RDONLY) = -1 ENOENT (No such file or directory)
[pid 13996] stat("/var/tmp", {st_mode=S_IFDIR|0755, st_size=40, ...}) = 0
[pid 13996] open("/var/lib/rpm/__db.001", O_RDWR|O_CREAT|O_EXCL, 0644) = -1 EEXIST (File exists)
[pid 13996] open("/var/lib/rpm/__db.001", O_RDWR) = 10
[pid 13996] fcntl(10, F_GETFD)          = 0
[pid 13996] fcntl(10, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] flock(10, LOCK_SH)          = 0
[pid 13996] fstat(10, {st_mode=S_IFREG|0644, st_size=352256, ...}) = 0
[pid 13996] read(10, "\227\10\22\0\0\0\0\0\5\0\0\0\3\0\0\0\34\0\0\0q\"h\2\241 \356\207\0\0\0\0"..., 168) = 168
[pid 13996] lseek(10, 480, SEEK_SET)    = 480
[pid 13996] read(10, "\0`\5\0\0\0\0\0\0\300\17\0\0\0\0\0\377\377\377\377\377\377\377\377", 24) = 24
[pid 13996] open("/var/lib/rpm/__db.001", O_RDWR) = 11
[pid 13996] fcntl(11, F_GETFD)          = 0
[pid 13996] fcntl(11, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] mmap(NULL, 1032192, PROT_READ|PROT_WRITE, MAP_SHARED, 11, 0) = 0x7f2296279000
[pid 13996] stat("/lib64/libpthread.so.0", {st_mode=S_IFREG|0755, st_size=142736, ...}) = 0
[pid 13996] open("/var/lib/rpm/__db.002", O_RDWR|O_CREAT, 0644) = 12
[pid 13996] fcntl(12, F_GETFD)          = 0
[pid 13996] fcntl(12, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] mmap(NULL, 262144, PROT_READ|PROT_WRITE, MAP_SHARED, 12, 0) = 0x7f2296be4000
[pid 13996] open("/var/lib/rpm/__db.003", O_RDWR|O_CREAT, 0644) = 17
[pid 13996] fcntl(17, F_GETFD)          = 0
[pid 13996] fcntl(17, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] mmap(NULL, 1318912, PROT_READ|PROT_WRITE, MAP_SHARED, 17, 0) = 0x7f229225e000
[pid 13996] close(8)                    = 0
[pid 13996] stat("/var/lib/rpm/Packages", {st_mode=S_IFREG|0644, st_size=12288, ...}) = 0
[pid 13996] stat("/var/lib/rpm/Packages", {st_mode=S_IFREG|0644, st_size=12288, ...}) = 0
[pid 13996] open("/var/lib/rpm/Packages", O_RDONLY) = 8
[pid 13996] fcntl(8, F_GETFD)           = 0
[pid 13996] fcntl(8, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] read(8, "\0\0\0\0\1\0\0\0\0\0\0\0a\25\6\0\t\0\0\0\0\20\0\0\0\10\0\0\0\0\0\0"..., 512) = 512
[pid 13996] close(8)                    = 0
[pid 13996] stat("/var/lib/rpm/Packages", {st_mode=S_IFREG|0644, st_size=12288, ...}) = 0
[pid 13996] open("/var/lib/rpm/Packages", O_RDONLY) = 8
[pid 13996] fcntl(8, F_GETFD)           = 0
[pid 13996] fcntl(8, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] fstat(8, {st_mode=S_IFREG|0644, st_size=12288, ...}) = 0
[pid 13996] open("/var/lib/rpm/pubkeys", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = -1 ENOENT (No such file or directory)
[pid 13996] rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
[pid 13996] rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
[pid 13996] stat("/var/lib/rpm/Name", {st_mode=S_IFREG|0644, st_size=8192, ...}) = 0
[pid 13996] stat("/var/lib/rpm/Name", {st_mode=S_IFREG|0644, st_size=8192, ...}) = 0
[pid 13996] open("/var/lib/rpm/Name", O_RDONLY) = 18
[pid 13996] fcntl(18, F_GETFD)          = 0
[pid 13996] fcntl(18, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] read(18, "\0\0\0\0\1\0\0\0\0\0\0\0b1\5\0\t\0\0\0\0\20\0\0\0\t\0\0\0\0\0\0"..., 512) = 512
[pid 13996] close(18)                   = 0
[pid 13996] stat("/var/lib/rpm/Name", {st_mode=S_IFREG|0644, st_size=8192, ...}) = 0
[pid 13996] open("/var/lib/rpm/Name", O_RDONLY) = 18
[pid 13996] fcntl(18, F_GETFD)          = 0
[pid 13996] fcntl(18, F_SETFD, FD_CLOEXEC) = 0
[pid 13996] fstat(18, {st_mode=S_IFREG|0644, st_size=8192, ...}) = 0
[pid 13996] read(7, "\355\253\356\333\3\0\0\0\0\23lib64attr1-2.4.48-1.mg"..., 96) = 96
[pid 13996] read(7, "\216\255\350\1\0\0\0\0\0\0\0\t\0\0\20\224", 16) = 16
[pid 13996] read(7, "\0\0\0>\0\0\0\7\0\0\20\204\0\0\0\20\0\0\1\f\0\0\0\7\0\0\0\0\0\0\2\30"..., 4388) = 4388
[pid 13996] read(7, "\0\0\0\0", 4)      = 4
[pid 13996] read(7, "\216\255\350\1\0\0\0\0\0\0\0=\0\0\10T", 16) = 16
[pid 13996] read(7, "\0\0\0?\0\0\0\7\0\0\10D\0\0\0\20\0\0\0d\0\0\0\10\0\0\0\0\0\0\0\1"..., 3108) = 3108
[pid 13996] close(7)                    = 0
[pid 13996] close(8)                    = 0
[pid 13996] close(18)                   = 0
[pid 13996] umask(022)                  = 022
[pid 13996] open("/var/lib/rpm/.dbenv.lock", O_RDWR|O_CREAT, 0644) = 7
[pid 13996] umask(022)                  = 022
[pid 13996] fcntl(7, F_SETLKW, {l_type=F_WRLCK, l_whence=SEEK_SET, l_start=0, l_len=0}) = 0
[pid 13996] close(17)                   = 0
[pid 13996] munmap(0x7f229225e000, 1318912) = 0
[pid 13996] close(12)                   = 0
[pid 13996] munmap(0x7f2296be4000, 262144) = 0
[pid 13996] close(10)                   = 0
[pid 13996] close(11)                   = 0
[pid 13996] munmap(0x7f2296279000, 1032192) = 0
[pid 13996] close(7)                    = 0
[pid 13996] rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1], [], 8) = 0
[pid 13996] rt_sigaction(SIGINT, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, NULL, 8) = 0
[pid 13996] rt_sigaction(SIGQUIT, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, NULL, 8) = 0
[pid 13996] rt_sigaction(SIGHUP, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, NULL, 8) = 0
[pid 13996] rt_sigaction(SIGTERM, {sa_handler=SIG_DFL, sa_mask=[], sa_flags=SA_RESTORER, sa_restorer=0x7f22a8ffea70}, NULL, 8) = 0
[pid 13996] rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
[pid 13996] lstat("/mnt/var/cache/urpmi/partial/lib64attr1-2.4.48-1.mga7.x86_64.rpm", {st_mode=S_IFREG|0644, st_size=15768, ...}) = 0
[pid 13996] unlink("/mnt/var/cache/urpmi/partial/lib64attr1-2.4.48-1.mga7.x86_64.rpm") = 0

Comment 60 Frank Griffin 2018-01-21 00:54:56 CET

Perhaps open the strace to all I/O and see what precedes the stderr ?

Comment 61 Pascal Terjan 2018-01-21 00:57:58 CET

So the code is verify_partial_rpm_and_move, I'll try to investigate URPM::verify_rpm:

sub verify_partial_rpm_and_move {
    my ($urpm, $cachedir, $filename) = @_;

    URPM::verify_rpm("$cachedir/partial/$filename", nosignatures => 1) or do {
        unlink "$cachedir/partial/$filename";
        return;
    };
    #- it seems the the file has been downloaded correctly and has been checked to be valid.
    unlink "$cachedir/rpms/$filename";
    urpm::sys::move_or_die($urpm, "$cachedir/partial/$filename", "$cachedir/rpms/$filename");
    "$cachedir/rpms/$filename";
}

Comment 62 Pascal Terjan 2018-01-21 01:11:14 CET

OK from inside installer root:

[root@mageia /]# perl -mURPM -e 'URPM::verify_rpm("/mnt/var/cache/urpmi/partial/lib64brotlicommon1-1.0.2-1.mga7.x86_64.rpm", nosignatures=>1) or print"bad\n";'
bad

From outside:
[pterjan@mageia git]$ perl -mURPM -e 'URPM::verify_rpm("/tmp/drakx-in-chroot/mnt/var/cache/urpmi/partial/lib64brotlicommon1-1.0.2-1.mga7.x86_64.rpm", nosignature=>1) or print "bad\n";'
[pterjan@mageia git]$

Comment 63 Pascal Terjan 2018-01-21 01:34:29 CET

After patching URPM.xs to not hide rpm error message:

# chroot /tmp/drakx-in-chroot/ perl -mURPM -e 'URPM::verify_rpm("/mnt/var/cache/urpmi/partial/lib64curl4-7.57.0-2.mga7.x86_64.rpm", nosignatures=>1, nodigests=>1) or print"bad\n";'
/mnt/var/cache/urpmi/partial/lib64curl4-7.57.0-2.mga7.x86_64.rpm: digests SIGNATURES NOT OK
bad

Comment 64 Pascal Terjan 2018-01-21 01:36:55 CET

So it seems nosignatures is not longer working

Comment 65 Pascal Terjan 2018-01-21 02:20:57 CET

This is the commit which broke it https://github.com/rpm-software-management/rpm/commit/80f8d6376066a3e2e303965d7f0e462d0f2bb8bb#diff-0fddb6fc4f41d1373a1860b54cda7ae0

Comment 66 Pascal Terjan 2018-01-21 02:27:03 CET

this is quite ugly but rpmVerifySignatures passes qva_flags as rpmVSFlags to rpmpkgVerifySigs so we could just use the new value, I'll try to patch URPM.

Comment 67 Pascal Terjan 2018-01-21 02:33:23 CET

This works, but breaks compatibility with older rpm:

diff --git a/URPM.xs b/URPM.xs
index 65729a5..88864ce 100644
--- a/URPM.xs
+++ b/URPM.xs
@@ -3080,15 +3080,15 @@ Urpm_verify_rpm(filename, ...)
   oldlogmask = rpmlogSetMask(RPMLOG_UPTO(RPMLOG_PRI(4)));
   memset(&qva, 0, sizeof(struct rpmQVKArguments_s));
   qva.qva_source = RPMQV_RPM;
-  qva.qva_flags = VERIFY_ALL;
+  qva.qva_flags = RPMVSF_DEFAULT;
   for (i = 1 ; i < items - 1 ; i += 2) {
     STRLEN len;
     char *s = SvPV(ST(i), len);
     if (SvIV(ST(i+1))) {
       if (len == 9 && !strncmp(s, "nodigests", 9))
-        qva.qva_flags &= ~VERIFY_DIGEST;
+        qva.qva_flags |= _RPMVSF_NODIGESTS;
       else if (len == 12 && !strncmp(s, "nosignatures", 12))
-        qva.qva_flags &= ~VERIFY_SIGNATURE;
+        qva.qva_flags |= _RPMVSF_NOSIGNATURES;
     }
   }
   fd = Fopen(filename, "r");

I think the cleaner way would be to stop using rpmVerifySignatures as it is broken by new rpm totally changing the semantic of qva_flags passed to it

Comment 68 Mageia Robot 2018-01-21 13:36:37 CET

commit 9f3525d0345c1f4e91c5a7a68043e8efe0a2b5b5
Author: Pascal Terjan <pterjan@...>
Date:   Sun Jan 21 12:31:14 2018 +0000

    Fix Urpm_verify_rpm during install
    
    Signatures can't be verified as pubkeys are not available, the code
    ignores it but that broke with rpm 4.14 (mga#21886).
    Switch to using rpmtsSetVSFlags and rpmReadPackageFile which should
    work with older versions too.
---
 Commit Link:
   http://gitweb.mageia.org/software/rpm/perl-URPM/commit/?id=9f3525d0345c1f4e91c5a7a68043e8efe0a2b5b5

Comment 69 Pascal Terjan 2018-01-21 14:22:52 CET

Assigning to Thierry to decide if it's in a state to be released

Assignee: mageiatools => thierry.vignaud

Comment 70 Thierry Vignaud 2018-01-21 15:30:22 CET

I could do a test upload to CPAN in order to check that it still work on a wide array of systems:
http://matrix.cpantesters.org/?dist=URPM%205.15.14
But I don't think URPM testsuite covers signature checking...

If that works fine when compiling+testing on both mga6 & cauldron, I guess you're good to go.
Can you check the same code works as before on mga6?
Thx

Thierry Vignaud 2018-01-21 15:30:35 CET

Source RPM: drakx-installer-stage2 => drakx-installer-stage2, rpm-4.14

Thierry Vignaud 2018-01-21 15:30:40 CET

Source RPM: drakx-installer-stage2, rpm-4.14 => drakx-installer-stage2, rpm-4.14, perl-URPM

Comment 71 Thierry Vignaud 2018-01-21 15:31:27 CET

Though, the best course of action would be to download the keys & check them... (security...)

Comment 72 Pascal Terjan 2018-01-21 18:07:09 CET

I first thought of embedding the pubkeys in the stage2 image, but given that it's likely to come from the same mirror the stage2 was downloaded from, that doesn't help much as the code to check them could be removed...

It seems other changes break on Mageia 6:

URPM.xs: In function 'XS_URPM_parse_rpm':
URPM.xs:3028:31: error: 'RPMVSF_NOPAYLOAD' undeclared (first use in this function)
                   vsflags |= (RPMVSF_NOPAYLOAD | RPMVSF_NOSHA1HEADER);
                               ^
URPM.xs:3028:31: note: each undeclared identifier is reported only once for each function it appears in
URPM.xs:3032:46: error: 'RPMVSF_NOSHA256HEADER' undeclared (first use in this function)
                   vsflags |= (RPMVSF_NOMD5 | RPMVSF_NOSHA256HEADER);
                                              ^
URPM.xs: In function 'XS_URPM_get_gpg_fingerprint':
URPM.xs:3121:9: warning: implicit declaration of function 'pgpPubkeyKeyID' [-Wimplicit-function-declaration]
         pgpPubkeyKeyID (pkt, pktlen, fingerprint);
         ^
URPM.c: In function 'XS_URPM__Package_EVR':
URPM.xs:1504:10: warning: ignoring return value of 'asprintf', declared with attribute warn_unused_result [-Wunused-result]
          asprintf(&res, "%d:%s", epoch, version);
          ^
Makefile:349: recipe for target 'URPM.o' failed

Comment 73 Dave Hodgins 2018-01-21 22:37:34 CET

To be as secure as we can make it, the keys would have to be embedded in stage1,
which downloaders can verify using the published sha512 check sums.

Comment 74 Frank Griffin 2018-02-03 16:15:26 CET

Can one of the other verifiers please try this again ?  When I try now I find X locks up immediately when we hit stage 2.  I'd be interested to know if that's just me.

Comment 75 Bit Twister 2018-02-03 17:24:59 CET

(In reply to Frank Griffin from comment #74)
> Can one of the other verifiers please try this again ? 

I used grub2 to boot Feb  1 00:59 Mageia-Cauldron-netinstall-nonfree-x86_64.iso
from my hard drive and get the first screen, and mouse will not move.
Ctrl+Delete or Backspace do nothing. Had to cold boot system.

CC: (none) => bittwister2

Comment 76 Thomas Bigot 2018-02-05 09:31:55 CET

Frozen screen here too.

Comment 77 Frank Griffin 2018-02-05 15:05:45 CET

See bug#22517 for the X hang.

Comment 78 Mike Rambo 2018-02-05 15:15:35 CET

Freezes (or at least no response to mouse/kybd) at the language selection screen. Tested in VB.

Comment 79 Thierry Vignaud 2018-02-09 17:32:37 CET

(In reply to Pascal Terjan from comment #65)
> This is the commit which broke it
> https://github.com/rpm-software-management/rpm/commit/
> 80f8d6376066a3e2e303965d7f0e462d0f2bb8bb#diff-
> 0fddb6fc4f41d1373a1860b54cda7ae0

The thing is that urpmi does import the pubkeys when adding the media, way before attempting to install the pkgs...
So there's sg fishy here

Comment 80 Thierry Vignaud 2018-02-09 17:34:58 CET

(In reply to Mike Rambo from comment #78)
> Freezes (or at least no response to mouse/kybd) at the language selection
> screen. Tested in VB.

It's "only" the input that's broken.
You can still got an automatic install with gfx being updated.

I suspect either a kernel HID regression or a xserver change, but the latest one there was for new monitor ratios.
@tmb: an idea?

Comment 81 Thierry Vignaud 2018-02-09 17:44:28 CET

& libinput of course

Comment 82 Thierry Vignaud 2018-02-09 17:55:38 CET

Created attachment 9970 [details]
xorg.log from an older install

Comment 83 Thierry Vignaud 2018-02-09 17:56:06 CET

Created attachment 9971 [details]
xorg.log from today install

the libinput xorg driver is no more loaded

Comment 84 Mageia Robot 2018-02-09 18:14:34 CET

commit 05c156148d7d2e91b541fa8da7123422fccc26ee
Author: Thierry Vignaud <thierry.vignaud@...>
Date:   Fri Feb 9 17:58:45 2018 +0100

    include libinput too
    
    helps mga#21886
---
 Commit Link:
   http://gitweb.mageia.org/software/drakx/commit/?id=05c156148d7d2e91b541fa8da7123422fccc26ee

Comment 85 Thierry Vignaud 2018-02-09 18:15:06 CET

And indeed "libinput list-devices" lists nothing in DrakX (unlike a real system)

Comment 86 Frank Griffin 2018-02-09 21:07:57 CET

Just tried with the newly-arrived installer images, but the problem remains.  Does this need a rebuild of mdkinst.sqfs as well ?

Comment 87 Thierry Vignaud 2018-02-09 21:42:32 CET

(In reply to Thierry Vignaud from comment #85)
> And indeed "libinput list-devices" lists nothing in DrakX (unlike a real
> system)

All udev entries lack ID_INPUT* fields

Comment 88 Mageia Robot 2018-02-09 21:58:55 CET

commit 2212956636b63bed6cbacb090632c5681a1f924f
Author: Thierry Vignaud <thierry.vignaud@...>
Date:   Fri Feb 9 21:53:42 2018 +0100

    add input rules from new systemd (mga#21886)
    
    new systemd broke badly input devices under X.org in Drakx...
---
 Commit Link:
   http://gitweb.mageia.org/software/drakx/commit/?id=2212956636b63bed6cbacb090632c5681a1f924f

Comment 89 Thierry Vignaud 2018-02-09 22:18:18 CET

And back to the main issue:
1) urpmi has installed the gpg pubkey when adding the media
2) rpm access /mnt/var/lib/rpm (in the /mnt chroot)
3) but when checking for sig, it looks for /var/lib/rpm (not in the chroot)
It was a long time since rpm had such an issue :-(

Comment 90 Thierry Vignaud 2018-02-09 23:34:49 CET

And we're back to https://bugs.mageia.org/show_bug.cgi?id=21886#c65, as verify_rpm() now checks sig due to rpm.org changes, and thus try to access rpmdb outside of chroot as verify_rpm() doesn't know about chroot prefix

Comment 91 Thierry Vignaud 2018-02-09 23:38:56 CET

@Pascal: did you try keeping the previous code and just change s/VERIFY_SIGNATURE/_RPMVSF_NOSIGNATURES/

Comment 92 Frank Griffin 2018-02-09 23:42:11 CET

OK, 22517 fixed, back to the main issue.

Comment 93 Thierry Vignaud 2018-02-10 10:31:08 CET

Rebuilding the images with the fix

Status: REOPENED => RESOLVED
Resolution: (none) => FIXED

Comment 94 Frank Griffin 2018-02-10 15:13:38 CET

It works !  Your children and your childrens' children wil rise up and call you blessed :-)