Bug 212781 (fc6longupgrade)
Summary: | Installer hangs on updating an existing installation | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Diego Padula <dialp> |
Component: | anaconda | Assignee: | Paul Nasrat <nobody+pnasrat> |
Status: | CLOSED RAWHIDE | QA Contact: | |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 6 | CC: | andrewz, dialp, joe, pdbrown, rmstevens |
Target Milestone: | --- | ||
Target Release: | --- | ||
Hardware: | i686 | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2007-05-21 19:43:08 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: | |||
Bug Depends On: | |||
Bug Blocks: | 150226 |
Description
Diego Padula
2006-10-29 01:45:53 UTC
Same here, tried to upgrade my notebook's fc5 to fc6. Reproducible: always Steps to Reproduce: 1. boot either using boot cd or rescue cd, no matter which boot arguments 2. use nfs or http as installation source 3. select "upgrade an existing FC" Actual results: Anaconda starts with a dependency check which takes about 20 minutes. A dialogue appears "click next to upgrade", then anaconda requests about 800 rpm files from the webserver and shows "installation process starting". 2 minutes later it shows "preparing transaction" and anaconda goes to 100% system load and does nothing. Tried it multiple times. After 120 minutes of melting my notebook with a 100% system load anaconda I gave up. The installer actually doesn't hang during dependency check, but the progress bar starts completing several minutes after dependency check has started, causing the impression that the installer isn't doing anything. After nearly 15 minutes the progress bar starts filling. The dependency check completes in about 20 minutes in my computer. So this is not a bug. Then it begins preparing transaction, but that progress bar is also completed. Can't reproduce what Phil's describing. However, when it is updating the first package (at the very begining of the progress bar) it switches to text mode and initiates a reboot. When the system is bakc on, you are feced again with the same install screen as if nothing had been done. If you remove the CD-ROM and boot the existing Fedora Core the system will fail to start X so previous Fedora Core installation is ruined. This looks like bug 208717, bug 208725, and bug 212575. I have the problem on two different Pentium 3 systems upgrading FC5->FC6 from different sets of verified boot media. The dependency check stays at 0% for a long time. Then, "Preparing transaction from" takes a long time and eventually fails in different ways (including a rpmdb panics like bug 211917). On one system, I killed it after waiting several hours. Both my systems became corrupt and the from-scratch installation hanged on package 6 (glibc-common) with 100% CPU usage. I am trying ide=nodma acpi=off apm=off and just got past glibc-common. :) I have not had such problems installing earlier versions of Fedora Core. I had the same problem I let it run all day and it took about 12 hours for the upgrade from FC5 to complete. I noticed that the installer was repeatedly calling ldconfig, and that may have been causing the install to run slow. *** Bug 213826 has been marked as a duplicate of this bug. *** After upgrading my notebook's memory from 256 MB to 512 MB the upgrade seems working fine. Still 30 mins to go %-) I also have 512MB ram and still had the problem. Perhaps, since I used a 64 bit install, I need even more ram? I appear to have the same issue. It has been sitting at "Preparing transaction from installation source...." for over 10 hours with no signs of moving at all. I thought to check whether it has perhaps run out of memory, so I ran 'cat /proc/meminfo' on a VT and it shows LowTotal: 385244 kB, LowFree: 5008 kB, SwapTotal: 313184 kB, SwapFree: 313184 kB. I'm not entirely sure what I'm looking at.... UPDATE: I ran strace on anaconda, and it appears to be hung giving the following message for the last 10 or 20 minutes straight: --- SIGSEGV (Segmentation fault) @ 0 (0) --- rt_sigaction(SIGSEGV, {0x8229f0, [], 0}, {0x8229f0, [], 0}, 8) = 0 sigreturn() = ? (mask now []) Things should be better here for Fedora 7. I've done a few low memory upgrades at least without problems As bug 253510 indicates, the problem remains. I think this bug should be REOPENED, but I don't have that access. Should we merge these two bugs? |