Hello,
about one hour ago, I did a "yum update" on a remote centos 4.3 server. It said:
Install 2 Package(s) Update 60 Package(s) Remove 0 Package(s) Total download size: 103 M Is this ok [y/N]: y
then, after downloading:
Running Transaction Test Finished Transaction Test Transaction Test Succeeded Running Transaction
Running Transaction Updating : libgcc ##################### [ 1/124] Updating : python-urlgrabber ##################### [ 2/124] Updating : hwdata ##################### [ 3/124]
and so on (note 124 packages, not 62), until it arrived at:
Editing /etc/postfix/master.cf, adding missing entry for tlsmgr service Updating : yum ##################### [ 62/124] Cleanup : coreutils ##################### [ 63/124] Cleanup : net-tools ##################### [ 64/124] Cleanup : tar ##################### [ 65/124] Cleanup : postfix ##################### [ 66/124] Cleanup : file ##################### [ 67/124] Cleanup : grep ##################### [ 68/124] Cleanup : newt ##################### [ 69/124] Cleanup : python-sqlite ##################### [ 70/124] Cleanup : rpm-build ##################### [ 71/124] Cleanup : mkinitrd ##################### [ 72/124] Cleanup : e2fsprogs ##################### [ 73/124]
and then nothing more. It has remained stuck at that line for almost one hour now. Is something funny happening? What should I do? Hit Control-C, halt the system from another shell (during an upgrade????)? or should I simply ignore it?
I have no problem to reinstall/reconfigure something if need be. I'm just worried to lose any contact with the server if I do the wrong thing during an apparently frozen update.
TIA, Marco
Marco,
Its in the cleanup stage, so at this point its removing the headers and rpms I believe. Your system is updated, and isnt getting any revelant information from the package servers. Someone correct me if I'm wrong.
Mike
On 9/11/06, M. Fioretti mfioretti@mclink.it wrote:
Hello,
about one hour ago, I did a "yum update" on a remote centos 4.3 server. It said:
Install 2 Package(s) Update 60 Package(s) Remove 0 Package(s) Total download size: 103 M Is this ok [y/N]: y
then, after downloading:
Running Transaction Test Finished Transaction Test Transaction Test Succeeded Running Transaction
Running Transaction Updating : libgcc ##################### [ 1/124] Updating : python-urlgrabber ##################### [ 2/124] Updating : hwdata ##################### [ 3/124]
and so on (note 124 packages, not 62), until it arrived at:
Editing /etc/postfix/master.cf, adding missing entry for tlsmgr service Updating : yum ##################### [ 62/124] Cleanup : coreutils ##################### [ 63/124] Cleanup : net-tools ##################### [ 64/124] Cleanup : tar ##################### [ 65/124] Cleanup : postfix ##################### [ 66/124] Cleanup : file ##################### [ 67/124] Cleanup : grep ##################### [ 68/124] Cleanup : newt ##################### [ 69/124] Cleanup : python-sqlite ##################### [ 70/124] Cleanup : rpm-build ##################### [ 71/124] Cleanup : mkinitrd ##################### [ 72/124] Cleanup : e2fsprogs ##################### [ 73/124]
and then nothing more. It has remained stuck at that line for almost one hour now. Is something funny happening? What should I do? Hit Control-C, halt the system from another shell (during an upgrade????)? or should I simply ignore it?
I have no problem to reinstall/reconfigure something if need be. I'm just worried to lose any contact with the server if I do the wrong thing during an apparently frozen update.
TIA, Marco _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
On Mon, Sep 11, 2006 12:34:58 PM -0700, mike opoien (mike.opoien@gmail.com) wrote:
Marco,
Its in the cleanup stage, so at this point its removing the headers and rpms I believe. Your system is updated,
OK, so if you are right I can safely reboot it without the risk that it cannot come up again, right?
For the record, it's still hanged at the same point.
Thanks,
Marco
Yeah I do not believe you are hosed. You could check you logs and see what kind of errors are throwing, but I dont believe you have much of a choice in the matter.
On 9/11/06, M. Fioretti mfioretti@mclink.it wrote:
On Mon, Sep 11, 2006 12:34:58 PM -0700, mike opoien (mike.opoien@gmail.com) wrote:
Marco,
Its in the cleanup stage, so at this point its removing the headers and rpms I believe. Your system is updated,
OK, so if you are right I can safely reboot it without the risk that it cannot come up again, right?
For the record, it's still hanged at the same point.
Thanks,
Marco _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Le lundi 11 septembre 2006 21:26, M. Fioretti a écrit :
Hello,
about one hour ago, I did a "yum update" on a remote centos 4.3 server. It said:
[...]
Cleanup : e2fsprogs ##################### [ 73/124]
and then nothing more. It has remained stuck at that line for almost one hour now. Is something funny happening? What should I do? Hit Control-C, halt the system from another shell (during an upgrade????)? or should I simply ignore it?
I have no problem to reinstall/reconfigure something if need be. I'm just worried to lose any contact with the server if I do the wrong thing during an apparently frozen update.
TIA, Marco
erm, well, I got that too (freeze of yum update in cleanup stage). When I finally went to the console, the server was spitting lots of ext3 (filesystem) error messages :-( I had no other option than restart the barbarian way (I had the magic SysRq disabled, so maybe you can avoid that), the filesystem check seemed to do his job, but then I had a number of duplicate packages, which I cleaned using that command given by Scott Silva on september 1st in "4.4 yum update problem" thread:
for file in `rpm -qa --queryformat="%{NAME} %{ARCH}\n" | sort | uniq -c | grep -v " 1 " | cut -c 9- | cut -d" " -f1`; do rpm -q --last $file | tail -1 | cut -d" " -f1; done | grep -v gpg-pubkey | xargs rpm -e --justdb --nodeps 2> /tmp/dupes
(all on one line)
So I'd say check if there are messages on the console, and wait a little for comments from people in the know to see if there are better advices...
On Mon, Sep 11, 2006 22:33:36 PM +0200, Vincent Knecht (vknecht@club-internet.fr) wrote:
erm, well, I got that too (freeze of yum update in cleanup stage). When I finally went to the console, the server was spitting lots of ext3 (filesystem) error messages :-( I had no other option than restart the barbarian way (I had the magic SysRq disabled, so maybe you can avoid that),
SOrry, what is, in this case, the "barbarian way"? This is a remote server on UM Linux. What I can do is to log in at the emergency console, or whatever the ISP calls it, and type a halt command there. Is this what you are suggesting?
Also, I don't understand the "magic SysRq disabled part. What do you mean?
So I'd say check if there are messages on the console
typing C-z, C-c, what?
THanks, Marco
Le lundi 11 septembre 2006 22:50, M. Fioretti a écrit :
On Mon, Sep 11, 2006 22:33:36 PM +0200, Vincent Knecht
(vknecht@club-internet.fr) wrote:
erm, well, I got that too (freeze of yum update in cleanup stage). When I finally went to the console, the server was spitting lots of ext3 (filesystem) error messages :-( I had no other option than restart the barbarian way (I had the magic SysRq disabled, so maybe you can avoid that),
SOrry, what is, in this case, the "barbarian way"?
It was simply turning the machine off using the power switch, then power it on.
This is a remote server on UM Linux. What I can do is to log in at the emergency console, or whatever the ISP calls it, and type a halt command there. Is this what you are suggesting?
Well, I couldn't log in in any way (via console or sshd), so I had no other choice. In your case you may have other options since you seem to be able to somehow log in. At any rate, if the system just sits there doing nothing, you'll have to interrupt it one way or the other...
Also, I don't understand the "magic SysRq disabled part. What do you mean?
It's the kernel.sysrq setting as seen in /etc/sysctl.conf, which might work to send some last-resort command to not-so-uglily shutdown a machine even when Ctrl-Alt-Suppr doesn't work. I'm not sure that applies to your setup, I never used nor really learnt about UM Linux (User Mode Linux I guess).
So I'd say check if there are messages on the console
typing C-z, C-c, what?
I just got error messages scrolling over all virtual consoles, without typing anything. Again, I don't know if something special should/could be done given your setup. If console shows nothing special, I'd say check /var/log/yum.log, /var/log/messages and "dmesg" command output.
On Mon, Sep 11, 2006 21:26:19 PM +0200, io (mfioretti@mclink.it) wrote:
about one hour ago, I did a "yum update" on a remote centos 4.3 server. It said:
Cleanup : mkinitrd ##################### [ 72/124] Cleanup : e2fsprogs ##################### [ 73/124]
and then nothing more. It has remained stuck at that line for almost one hour now.
Hello, it's me again.
I don't know if it's relevant, but I have connected via ssh as normal user without seeing any error. Then, when I switch to root, I get:
$ su - -bash: child setpgid (9021 to 9021): No such process Password:
I have seen via Google that others, to fix these warnings, suggested to:
"mv /lib/tls /tls-disabled as noted in the 2.6 kernel announcements."
No problem from me to do that. Just dubious if (considering the hanged update) it is something to do before, after, instead typing "halt" at this second ssh window I have.
Thoughts, comments, explanations of what is actually happening?
Thanks in advance for any feedback,
Marco
M. Fioretti spake the following on 9/11/2006 1:58 PM:
On Mon, Sep 11, 2006 21:26:19 PM +0200, io (mfioretti@mclink.it) wrote:
about one hour ago, I did a "yum update" on a remote centos 4.3 server. It said:
Cleanup : mkinitrd ##################### [ 72/124] Cleanup : e2fsprogs ##################### [ 73/124]
and then nothing more. It has remained stuck at that line for almost one hour now.
Hello, it's me again.
I don't know if it's relevant, but I have connected via ssh as normal user without seeing any error. Then, when I switch to root, I get:
$ su - -bash: child setpgid (9021 to 9021): No such process Password:
I have seen via Google that others, to fix these warnings, suggested to:
"mv /lib/tls /tls-disabled as noted in the 2.6 kernel announcements."
No problem from me to do that. Just dubious if (considering the hanged update) it is something to do before, after, instead typing "halt" at this second ssh window I have.
Thoughts, comments, explanations of what is actually happening?
Thanks in advance for any feedback,
Marco
If you can't get another root prompt in another session, then a reboot might be your only choice. IF the system seems to be responding otherwise, you could take your time and reboot. If system is not responding, then reboot is probably the only choice. You don't want to send a "halt" unless you have some way to turn the system back on remotely.
-
M. Fioretti -
mike opoien -
Scott Silva -
Vincent Knecht