Bug 472933
Summary: | Preupgrade-cli not working with static IP assignment | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Jon Baker <jbaker> |
Component: | preupgrade | Assignee: | Seth Vidal <skvidal> |
Status: | CLOSED NEXTRELEASE | QA Contact: | Fedora Extras Quality Assurance <extras-qa> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 9 | CC: | bdwheele, chris, jbaker, misek, raytodd, wwoods |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2009-01-07 09:08:26 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: |
Description
Jon Baker
2008-11-25 17:36:25 UTC
P.S. - this is the machine that is serving DHCP, so if it's trying to get an IP address via DHCP (by either ignoring the options or having them set incorrectly) obviously it won't be successful. On the server at the server farm of course there is no DHCP, which is the whole purpose of this test. Further testing: I installed a minimal installation of FC9 on an independent box. I ran preupgrade-cli using a similar command as above (changing the static IP address to a different address than what DHCP is serving it). Upon reboot, I received the same message "There was an error configuring your network interface." I then ran preupgrade-cli --vnc=[vnc_password] "Fedora 10 (Cambridge)" to see if it would work with DHCP. Again, I recived the message "There was an error configuring your network interface." Then I ran preupgrade-cli "Fedora 10 (Cambridge)" to leave everything at the defaults. Now, on this other box there was enough room to download the install image so it technically didn't need to connect to the internet (I assume) - however, I got a different error this time. The process got to the blue screen and then said "install exited abnormally [0/0] -- received signal 9" - whether that's related or not I am not sure. Just for kicks I then ran preupgrade-cli --ip=172.22.12.250 --netmask=255.255.255.0 --gateway=172.22.12.1 --dns=172.22.12.1 "Fedora 10 (Cambridge)" That got me the "install exited abnormally" error message again. I'm guessing that without the vnc option, it's not trying to activate the network card; neither IP address responded to any pings during the reboot. Hmm the same in my case, I run preupgrade-cli "Fedora 10 (Cambridge)" and after reboot the install stopped with "There was an error configuring your network interface". In F9 there was NM used with two static interfaces. I'm planning to connect the machine to the net with DHCP if it makes any difference. I had the same problem. The work around is to modify the /boot/upgrade/ks.cfg file by adding a line like this after the "lang" line: network --bootproto static --ip xxx.xxx.xxx.xxx --netmask xxx.xxx.xxx.xxx -gateway xxx.xxx.xxx.xxx --nameserver xxx.xxx.xxx.xxx After the reboot it worked ok. I wasn't using VNC, so your mileage may vary. (In reply to comment #4) > I had the same problem. The work around is to modify the /boot/upgrade/ks.cfg > file by adding a line like this after the "lang" line: > > network --bootproto static --ip xxx.xxx.xxx.xxx --netmask xxx.xxx.xxx.xxx > -gateway xxx.xxx.xxx.xxx --nameserver xxx.xxx.xxx.xxx > > > After the reboot it worked ok. I wasn't using VNC, so your mileage may vary. thanks for the tip, that did get me to the point where it initializes the network and completes the download of the install.img file. Next it boots into anaconda, but asks me to enter a vnc password - which was set in the grub.conf and obviously can't happen on a headless system. Did everyone here also get the message about /boot being too small? This only shows up when the URL for stage2=XXX is really long. As far as I can tell, the problem is that the commandline gets too long and the network configuration parameters are being cut off. That's why moving them into ks.cfg fixes the problem. Maybe that's the right solution to the "I need network" problem. Oh, by the way - the commandline-network-setup stuff is only supposed to be used for VNC (remote, headless) installs. If you have local access to the machine you can just set up the network in the menus once it starts the upgrade. Yes, I did (In reply to comment #6) > Did everyone here also get the message about /boot being too small? Yes, I did - > This only shows up when the URL for stage2=XXX is really long. As far as I can > tell, the problem is that the commandline gets too long and the network > configuration parameters are being cut off. > > That's why moving them into ks.cfg fixes the problem. Maybe that's the right > solution to the "I need network" problem. That makes sense, maybe the app needs to put the info into ks.cfg as well as the grub.conf file? > Oh, by the way - the commandline-network-setup stuff is only supposed to be > used for VNC (remote, headless) installs. If you have local access to the > machine you can just set up the network in the menus once it starts the > upgrade. Yes, that's what I'm trying to do - only I'm testing it on a local system before trying it on a headless system, since it's much more complicated to get a headless system back when it's in a data center across the country and you're at the mercy of the tech as to when it can be rebooted; just wanted to actually see what it's going to do with the box across the room before trying it on a box across the country. Hey Will, I think you're right. I got it working by doing this: 1) Run the preupgrade-cli with just the vnc option 2) Grab the stage2 URL and send it through tinyurl.com to get a small URL 3) Replace the stage2 value with the new URL, delete the ip info in the command 4) Set up ks.cfg as described above to activate the network This time it rebooted and launched the VNC server for the upgrade. Yeah, I can reproduce this by setting up a system with a small (or full) /boot partition and attempting a vnc-based install. The network options get cut off - you can check /tmp/anaconda.log during the install and see the truncated commandline there. For F11 I should be able to get rid of some of the really long URLs and UUIDs on the commandline, but for now I think putting the network/vnc config in the kickstart file is the right way to go. I'm working on an updated version of preupgrade that does that, so I may ask for testers soon. As a side note - I thought the kernel commandline was limited to 512 characters, but these commandlines are only around 350. So either I was wrong about the kernel limit or something else (GRUB? Anaconda?) is truncating the commandline. I'll investigate that further and file a new bug if necessary. *** Bug 475027 has been marked as a duplicate of this bug. *** I don't know if 475037 is a duplicate of this or not. Certainly there are some serious differences in the presentation of the bug if this is a duplicate. But they are close enough that they might be the same thing, and either way it might be good to work on them together. This bug seems to deal with the command line version of preupgrade not obeying information provided about the static ip. The bug I filed is for the console based version, where there is no way to even attempt to enter this data. These setting should be used from the old instalation as far as I understand preupgrade tool, so no network informations should be entered after preupgrade is finished (during anaconda state). This is definitely not what happens. Even if the previous installation has static information in place it tries to do dhcp and if the dhcp fails it just hangs up. This was posted to my bug marked as duplicate after it was closed so I am moving it here. Comment #2 From Will Woods (wwoods) 2008-12-15 13:33:28 EDT (-) [reply] ------- Why do you need the network configured at all during the upgrade? Did you get the "/boot too small" message? First I don't get the boot to small message, but it does normally try to store a file several times the size of the 100 meg boot partition in the boot partition and fails. Even if that is not the case it appears to try and check for updates as a part of the install process. Either way no network means that the upgrade fails. The problem in bug 475027 (And comment #11 through comment #14) is different from this one. I've reopened that bug. This bug is for trouble with preupgrade-cli and static networking, due to the boot commandline growing too long. Bug 475027 is for problems with preupgrade (the normal GUI) and non-DHCP networks. Please don't mix the two. Works for me. OK my comment 16 was in retrospect a little less than specific. I was not referencing that the program is working for me, but instead that the comment 15 from Will was acceptable to me and that I was moving my discussion of this subject back to that bug. Getting back to the original bug - preupgrade-1.0.1 has been built, which shortens the commandline considerably *and* puts the network configuration done in preupgrade-cli into the kickstart file. http://koji.fedoraproject.org/koji/buildinfo?buildID=73594 I've tested it a few times around the lab and it seems to work fine for me. If anyone can retest with that package and confirm, that would be great. If I don't hear anything bad I'll be pushing an update out soon. preupgrade-1.0.1-1.fc8 has been submitted as an update for Fedora 8. http://admin.fedoraproject.org/updates/preupgrade-1.0.1-1.fc8 preupgrade-1.0.1-1.fc9 has been submitted as an update for Fedora 9. http://admin.fedoraproject.org/updates/preupgrade-1.0.1-1.fc9 preupgrade-1.0.1-1.fc10 has been submitted as an update for Fedora 10. http://admin.fedoraproject.org/updates/preupgrade-1.0.1-1.fc10 preupgrade-1.0.1-1.fc8 has been pushed to the Fedora 8 testing repository. If problems still persist, please make note of it in this bug report. If you want to test the update, you can install it with su -c 'yum --enablerepo=updates-testing-newkey update preupgrade'. You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F8/FEDORA-2008-11587 preupgrade-1.0.1-1.fc9 has been pushed to the Fedora 9 testing repository. If problems still persist, please make note of it in this bug report. If you want to test the update, you can install it with su -c 'yum --enablerepo=updates-testing-newkey update preupgrade'. You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F9/FEDORA-2008-11540 preupgrade-1.0.1-1.fc10 has been pushed to the Fedora 10 testing repository. If problems still persist, please make note of it in this bug report. If you want to test the update, you can install it with su -c 'yum --enablerepo=updates-testing update preupgrade'. You can provide feedback for this update here: http://admin.fedoraproject.org/updates/F10/FEDORA-2008-11437 Just ran preupgrade-cli 1.0.1-1.fc8 on one of my servers which required the disk image download. I did get the network initialization error but when I hit retry it succeeded. Not sure exactly why that happened other than this is a very old box and we've had problems with the network card initializing in even FC6. But other than that, it worked perfectly. Thanks for the fix! preupgrade-1.0.1-1.fc8 has been pushed to the Fedora 8 stable repository. If problems still persist, please make note of it in this bug report. preupgrade-1.0.1-1.fc9 has been pushed to the Fedora 9 stable repository. If problems still persist, please make note of it in this bug report. preupgrade-1.0.1-1.fc10 has been pushed to the Fedora 10 stable repository. If problems still persist, please make note of it in this bug report. |