[Solved] Upgrade to pacman-4.0.3-2 fails

djraymondnm · 2012-06-05 16:02:20

After successfully upgrading to pacman-4.0.3-2 on 3 systems, I am getting a failure on the 4th:

[root@peregrine etc]# pacman -Suy
:: Synchronizing package databases...
core is up to date
extra is up to date
community is up to date
:: The following packages should be upgraded first :
pacman
:: Do you want to cancel the current operation
:: and upgrade these packages now? [Y/n] y

resolving dependencies...
looking for inter-conflicts...

Targets (1): pacman-4.0.3-2

Total Installed Size: 3.27 MiB
Net Upgrade Size: -0.27 MiB

Proceed with installation? [Y/n] y
(1/1) checking package integrity [######################] 100%
error: pacman: key "E62F853100F0D0F0" is unknown
:: Import PGP key 00F0D0F0, "Gaetan Bisson <g.bisson@tue.nl>", created 2010-01-09? [Y/n] y
(1/1) checking package integrity [######################] 100%
error: pacman: signature from "Gaetan Bisson <gaetan.bisson@normalesup.org>" is unknown trust
error: failed to commit transaction (invalid or corrupted package (PGP signature))
Errors occurred, no packages were upgraded.
[root@peregrine etc]#

I tried reverting to the -1 version of pacman, but no soap. Now I am stuck! How can I get pacman-4.0.3-2 installed successfully?

Last edited by djraymondnm (2012-06-08 02:43:53)

Gcool · 2012-06-05 16:21:40

What happens if you simply answer "N"?

Mr.Elendig · 2012-06-05 17:00:27

As a start, remove SyncFirst from pacman.conf, and then read this: http://www.archlinux.org/news/having-pa … -packages/

djraymondnm · 2012-06-05 18:03:51

1) Typing "N" just causes the install to abort.

2) Commenting out SyncFirst = pacman allows installation of other stuff to start, but it then aborts.
Running pacman-key --init simply hangs after the output "gpg: Generating pacman keychain master key..."

Pres · 2012-06-05 18:10:09

Read this to learn more about package signing:
https://wiki.archlinux.org/index.php/Package_signing

Alternately, just disable it if you do not want it. I'd suggest you take a couple minutes and learn about it/set it up though.

Terminator · 2012-06-05 18:11:59

Try disabling signature checking temporarily by setting 'SigLevel = Never' for all repositories in your pacman.conf. Then, you should be able to upgrade pacman and enable signature checking again.

edit: SigLevel = Never used to be the default, so I wonder what went wrong.

Last edited by Terminator (2012-06-05 18:13:04)

Inxsible · 2012-06-05 18:17:17

Terminator wrote:

edit: SigLevel = Never used to be the default, so I wonder what went wrong.

This

djraymondnm · 2012-06-05 18:32:18

The problem seems to be that pacman-key --init hangs on this particular machine. On my
other machines it finished really fast. Could there be some corrupt file standing in the way
of this command? I even tried removing /etc/pacman.d/gnupg and rerunning pacman-key --init.
It still hung.

Pres · 2012-06-05 18:39:30

Read the pacman-key page.

Terminator · 2012-06-05 18:43:47

Inxsible wrote:

Terminator wrote:
edit: SigLevel = Never used to be the default, so I wonder what went wrong.
This

I am aware of that. However, the OP says the upgrade to pacman-4.0.3-2 fails. Since this is the first version of pacman where signature checking is enabled by default, his pacman.conf should still contain 'SigLevel = Never'.

djraymondnm wrote:

The problem seems to be that pacman-key --init hangs on this particular machine. On my
other machines it finished really fast. Could there be some corrupt file standing in the way
of this command? I even tried removing /etc/pacman.d/gnupg and rerunning pacman-key --init.
It still hung.

Is the install on that machine one without a graphical environment? If so, the system might not find enough random data for pacman-key --init, move your mouse around; bash your keyboard; do what you need to do to give your pc enough data to initialize the keystore. Note that this happened to me a while ago, when I enabled the testing repositories before installing X. Also note that when that happened to me, running 'pacman-key --init' gave me comprehensible error-messages.

Last edited by Terminator (2012-06-05 19:03:09)

djraymondnm · 2012-06-05 20:04:42

OK, by setting "SigLevel = TrustAll" in pacman.conf, I can install things, including the new pacman.

However, doing pacman-key --init still hangs. Since pacman-key is a shell script (more or less),
I can trace the origin of the hang to the "%commit" statement in the routine "generate_master_key()"
in pacman-key. The "%commit" statement hangs. I can't trace this any further because I don't
know what "%commit" does.

Inxsible · 2012-06-05 20:11:02

Is the computer otherwise usable?

maybe its a case of

wiki wrote:

How can I collect entropy?
Moving your mouse around, pressing random characters at the keyboard or running some disk-based activity e.g. updatedb in another virtual console can solve it. It may take a while so please be patient.

djraymondnm · 2012-06-05 20:26:36

Found the problem: /etc/pacman.d/gnupg/secring.gpg was of zero length.
Copying the secring file from a healthy machine fixed the problem.

So how did this file get hosed???

doug piston · 2012-06-05 20:46:59

I deleted my /etc/pacman.d/gnupg as suggested here before upgrading pacman last night. Had no issues.

Edit: I'm not sure how that is relevant. Apologies. Moving on.

Last edited by doug piston (2012-06-05 20:47:54)

Mr.Elendig · 2012-06-06 11:16:43

djraymondnm wrote:

Found the problem: /etc/pacman.d/gnupg/secring.gpg was of zero length.
Copying the secring file from a healthy machine fixed the problem.
So how did this file get hosed???

Not the real problem.

You should just have been more patient, or used some of the fun tools to generate entropy. (network traffic works well too for it btw)

djraymondnm · 2012-06-07 03:33:45

Pacman-key --init finally works on the problem machine after about half an hour running lots of
"entropy generation" programs. (updatedb finally did the trick!)

This time is about 2 orders of magnitude larger than I experienced on any of 4 other machines.
Is it possible that some key generation situations just take much longer?

ewaller · 2012-06-07 03:47:15

djraymondnm wrote:

This time is about 2 orders of magnitude larger than I experienced on any of 4 other machines.

You are saying that the other machines finished on the order of 20 seconds. It is possible, especially if the machines had been on, doing other things prior to the key generation. The kernel does collect and store entropy continuously. Only when you run out do you need to wait. The system uses time between keystrokes, mouse movement, Ethernet latency, disk rotation latency, etc... It is dependent on the system components and whether they are running. Increasing numbers of systems even have hardware true random generators -- usually based on the metastability of a bi-stable circuit (flip-flops) or based upon jitter measurement of a "poor" oscillator. Maybe those other systems have some of those capabilities.

Is it possible that some key generation situations just take much longer?

You bet. Funny thing about entropy, it is not very organized

djraymondnm · 2012-06-08 02:50:03

ewaller wrote:

djraymondnm wrote:
This time is about 2 orders of magnitude larger than I experienced on any of 4 other machines.
You are saying that the other machines finished on the order of 20 seconds. It is possible, especially if the machines had been on, doing other things prior to the key generation. The kernel does collect and store entropy continuously. Only when you run out do you need to wait. The system uses time between keystrokes, mouse movement, Ethernet latency, disk rotation latency, etc... It is dependent on the system components and whether they are running. Increasing numbers of systems even have hardware true random generators -- usually based on the metastability of a bi-stable circuit (flip-flops) or based upon jitter measurement of a "poor" oscillator. Maybe those other systems have some of those capabilities.
Is it possible that some key generation situations just take much longer?
You bet. Funny thing about entropy, it is not very organized

I wonder if this information should go somewhere in the documentation.

ewaller · 2012-06-08 04:56:06

djraymondnm wrote:

[I wonder if this information should go somewhere in the documentation.

I must admit, I spend my time either here, hacking code, working, or spending time with my kids. Maybe this would give me an excuse to contribute to the wiki. The information I provided is not critical, but it does provide background.

Maybe this weekend ... After I pull the bearings on the washing machine, clean the seats, seat the new bearings, replace the seal, and re-assemble the piece of junk.
They say that being married to an engineer means you never get new appliances

whatshisname · 2012-06-17 23:10:49

djraymondnm wrote:

ewaller wrote:
djraymondnm wrote:
This time is about 2 orders of magnitude larger than I experienced on any of 4 other machines.
You are saying that the other machines finished on the order of 20 seconds. It is possible, especially if the machines had been on, doing other things prior to the key generation. The kernel does collect and store entropy continuously. Only when you run out do you need to wait. The system uses time between keystrokes, mouse movement, Ethernet latency, disk rotation latency, etc... It is dependent on the system components and whether they are running. Increasing numbers of systems even have hardware true random generators -- usually based on the metastability of a bi-stable circuit (flip-flops) or based upon jitter measurement of a "poor" oscillator. Maybe those other systems have some of those capabilities.
Is it possible that some key generation situations just take much longer?
You bet. Funny thing about entropy, it is not very organized
I wonder if this information should go somewhere in the documentation.

Probably a good idea because I just ran into the same problem. While my new Arch installed laptop was hung at "gpg: Generating pacman keychain master key...", I started researching the forums here and found this thread. While reading the thread, my laptop beeped and I had a prompt and, evidently, a new master key. Took a lot longer than it did on other systems.

Arch Linux

#1 2012-06-05 16:02:20

[Solved] Upgrade to pacman-4.0.3-2 fails

#2 2012-06-05 16:21:40

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#3 2012-06-05 17:00:27

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#4 2012-06-05 18:03:51

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#5 2012-06-05 18:10:09

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#6 2012-06-05 18:11:59

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#7 2012-06-05 18:17:17

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#8 2012-06-05 18:32:18

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#9 2012-06-05 18:39:30

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#10 2012-06-05 18:43:47

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#11 2012-06-05 20:04:42

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#12 2012-06-05 20:11:02

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#13 2012-06-05 20:26:36

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#14 2012-06-05 20:46:59

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#15 2012-06-06 11:16:43

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#16 2012-06-07 03:33:45

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#17 2012-06-07 03:47:15

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#18 2012-06-08 02:50:03

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#19 2012-06-08 04:56:06

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

#20 2012-06-17 23:10:49

Re: [Solved] Upgrade to pacman-4.0.3-2 fails

Board footer