PDA

View Full Version : [ubuntu] 9.04 upgrade via ssh... problem



unickque
December 20th, 2011, 08:54 AM
Hi all, i'm new to this forum and relatively new to Ubuntu, so please go easy on me :-)

I'm trying to do a remote upgrade of a 9.04 system. From the Ubuntu pages i found the only way to do so is to use the alternate cd. So I d/led the 9.10alternate_64.iso, mounted it, let it run, and it went okay, except after the reboot not many services were running (XDM and apache2 among others), but that's okay i thought, i'll fix that later.

After this initial success i went on to install 10.04 in the same way, and the success was repeated. I was on a roll.

Next the 10.10 install ran, and requested a reboot... only this time, the machine was no longer reachable after the reboot. I had to call someone on site, and he had to manually start the networking bit. After that I could login via ssh again, and it turned out there was an apt-get taking 100% cpu. I killed that process (it had been running over night, so i'm sure it should have been done). The system said it was running 10.10 Now... what to do i thought. I figured i'd rerun the 10.10 install just to make sure everything got installed. It did some stuff, and then asked for a reboot.

After some hesitation i figured i wouldn't get anywhere without a reboot, so i allowed it to reboot. But again, the system became unreachable.

I'll have to call the guy on site again today and make him bring up the network.

But the question is: what causes this problem? and what do I do next? Is it safe to run the 11.04 alternate install CD?

Please remember i have no local access to the machine, and XDM is not running right now so I have no graphics access yet.

Any advice is welcome.

mikewhatever
December 20th, 2011, 09:49 AM
It's hard to tell what exactly is the problem without deeper investigation. You should probably try checking the upgrade log to see if anything had gone wrong. Upgrading is never bulletproof, and upgrade sprees are probably even less so.

unickque
December 20th, 2011, 01:00 PM
I can post logs here if that is what you mean...
If someone can tell me where they are located?

matt_symes
December 20th, 2011, 01:22 PM
Hi

Upgrading via ssh ? You crazy nutter. :D

All your logs are kept in the directory /var/log. I would not bother upgrading again until you can track down why networking is not working after a reboot.

I assume you have not rebooted after the last time when you updated to 10.10 the second time ? It may be the case that if you reboot now then networking will restart after it.

However, unless you really want to irritate the person who has been restarting networking, i would try to track down what is happening to networking.

Have a look at the file


/var/log/syslog

and in


/var/log/installer/

on the remote machine.

I would not upgrade to 11.04 yet but i would look at the stability of the current system.

apt may have hung for a whole host of reasons.

As has been suggested, you need to investigate more.

Kind regards

unickque
December 20th, 2011, 01:35 PM
Hi

Upgrading via ssh ? You crazy nutter. :D

Well Ubuntu is used for servers afaik, so surely i'm not the first to do a remote upgrade like this?

If anyone has a better suggestion, please let me know! :) But the server is quite a distance away from my home, and this is only my hobby so i would have to take a day off for the installation. So going there myself is only my last option.

Thanks for pointing me to the logs, I'll see what I can find.

matt_symes
December 20th, 2011, 02:29 PM
Hi


Well Ubuntu is used for servers afaik, so surely i'm not the first to do a remote upgrade like this?

If anyone has a better suggestion, please let me know! :) But the server is quite a distance away from my home, and this is only my hobby so i would have to take a day off for the installation. So going there myself is only my last option.

Thanks for pointing me to the logs, I'll see what I can find.

Just to let you know that if you upgrade via ssh it will open another ssh connection in case you lose the first.

This is fine but not useful in you situation after a reboot.

Kind regards

unickque
December 20th, 2011, 06:47 PM
Nothing strange found so far.

In dmesg i find:

[ 1.128053] tg3 0000:03:01.0: eth0: Tigon3 [partno(N/A) rev 2100] (PCIX:133MHz:64-bit) MAC address 00:16:25:05:21:6d
[ 1.128058] tg3 0000:03:01.0: eth0: attached PHY is 5704 (10/100/1000Base-T Ethernet) (WireSpeed[1])
[ 1.128062] tg3 0000:03:01.0: eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[1] TSOcap[0]
[ 1.128065] tg3 0000:03:01.0: eth0: dma_rwctrl[769f4000] dma_mask[64-bit]
[ 1.128091] alloc irq_desc for 26 on node -1
[ 1.128094] alloc kstat_irqs on node -1
[ 1.128104] tg3 0000:03:01.1: PCI INT B -> GSI 26 (level, low) -> IRQ 26
[ 1.170222] cciss0: <0x46> at PCI 0000:04:03.0 IRQ 51 using DAC
[ 1.173525] cciss/c0d0: p1 p2 < p5 >
[ 1.191242] tg3 0000:03:01.1: eth1: Tigon3 [partno(N/A) rev 2100] (PCIX:133MHz:64-bit) MAC address 00:16:25:05:21:6c
[ 1.191248] tg3 0000:03:01.1: eth1: attached PHY is 5704 (10/100/1000Base-T Ethernet) (WireSpeed[1])
[ 1.191252] tg3 0000:03:01.1: eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1]
[ 1.191255] tg3 0000:03:01.1: eth1: dma_rwctrl[769f4000] dma_mask[64-bit]

...

[ 6.191274] type=1400 audit(1324313199.894:2): apparmor="STATUS" operation="profile_load" name="/sbin/dhclient3" pid=528 comm="apparmor_parser"
[ 6.191966] type=1400 audit(1324313199.894:3): apparmor="STATUS" operation="profile_load" name="/usr/lib/NetworkManager/nm-dhcp-client.action" pid=528 comm="apparmor_parser"
[ 6.192355] type=1400 audit(1324313199.894:4): apparmor="STATUS" operation="profile_load" name="/usr/lib/connman/scripts/dhclient-script" pid=528 comm="apparmor_parser"
[ 6.193247] type=1400 audit(1324313199.894:5): apparmor="STATUS" operation="profile_replace" name="/sbin/dhclient3" pid=531 comm="apparmor_parser"
[ 6.193263] type=1400 audit(1324313199.894:6): apparmor="STATUS" operation="profile_replace" name="/sbin/dhclient3" pid=512 comm="apparmor_parser"
[ 6.193954] type=1400 audit(1324313199.894:7): apparmor="STATUS" operation="profile_replace" name="/usr/lib/NetworkManager/nm-dhcp-client.action" pid=531 comm="apparmor_parser"
[ 6.193970] type=1400 audit(1324313199.894:8): apparmor="STATUS" operation="profile_replace" name="/usr/lib/NetworkManager/nm-dhcp-client.action" pid=512 comm="apparmor_parser"
[ 6.194374] type=1400 audit(1324313199.894:9): apparmor="STATUS" operation="profile_replace" name="/usr/lib/connman/scripts/dhclient-script" pid=512 comm="apparmor_parser"
[ 6.194394] type=1400 audit(1324313199.894:10): apparmor="STATUS" operation="profile_replace" name="/usr/lib/connman/scripts/dhclient-script" pid=531 comm="apparmor_parser"
[ 6.405307] ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 6.434463] input: PS/2 Generic Mouse as /devices/platform/i8042/serio1/input/input1
[ 6.816090] ADDRCONF(NETDEV_UP): eth1: link is not ready

I'm not sure whether this actually means the server is trying to use dhclient to get an ip address, but it shouldn't because it has a fixed IP. And I completely don't get why the network devices are reported 'not reay' ?

unickque
December 20th, 2011, 07:18 PM
Hmmm for some reason the resolv.conf was set aside and replaced by a new file containing only the line "search <domain.com>" by NetworkManager. As with most managers, he apparently didn't know what he was doing :P

Meanwhile the syslog shows that this same NetworkManager was looking for dns entries in /etc/network/interfaces

Did the resolv.conf get deprecated and replaced by some new method?

unickque
December 21st, 2011, 09:17 AM
Suppose i decide to just stay at the 10.10 version, how do i get the system up and running again after the alternate install?

Is there an apt-get install everything command?

I'll need php, apache, mysql at least, the rest can wait.


Also, is it possible to go back from 10.10 to 10.04? The LTS version might be a better option if the upgrade doesn't work?

Old_Grey_Wolf
December 21st, 2011, 09:47 PM
I'll need php, apache, mysql at least, the rest can wait.


Run these commands in the terminal to get Apache, mysql, php
sudo apt-get install tasksel

sudo tasksel install lamp-server

unickque
December 22nd, 2011, 08:29 AM
Okay thanks!

In the mean time i went on to 11.04 because i was sick of rebooting and losing contact with the server. That seemed to have helped a lot. The system rebooted okay and most things seemed to be running okay already. So i decided I like 11.04 and now i'm starting up all the services again. A few small hickups are still there, no secure websites yet, PHP5.3 is more picky than its predecessor, outgoing mail not working yet, .... but the websites are up and running so now i at least have everyone off my back and i can get back to normal problem solving :-)

Thanks for helping out everyone!