Red Hat Bugzilla – Bug 56948
repeated anaconda crashes during package install
Last modified: 2007-04-18 12:38:30 EDT
Description of Problem: Anaconda keeps crashing during the copy-and-install-packages portion of the install. It has crashed anywhere from 1/5th of the way done to 3/4 of the way done and seems unaffected by which packages I choose to install. This is with the "custom" install option. Version-Release number of selected component (if applicable): n/a How Reproducible: always Steps to Reproduce: 1. boot computer from bootnet.img 2. begin install from nfs share as usual. select custom (haven't tried other install options), autopartition, yada yada 3. sometime during the package install progress, anaconda crashes. I can still switch vts, use the shell etc. Actual Results: el crasho (see above). I am using slightly non-standard HW (a pc104 stack) but it's got a vanilla pentium 166 processor, normal-ish (128MB SODIMM) RAM, etc. Expected Results: A proper install would be the expected result. Additional Information: Here's a dump of what anaconda spews on crash. It's mostly the same each time (tried 4 times now) and I have noted in parentheses where the dumped data varies. unable to handle kernel paging request at virtual address edb794ad (the virtual address varies) printing eip: c01408e9 *pde = 00000000 Oops: 0002 CPU: 0 EIP: 0010:[<c01408e9>] (all that is always the same). EFLAGS: 00010a83 (sometimes 00010287) eax: edb794a9 ebx: c3fbd0b0 ecx: ddaf05c7 edx: c3fbd0b8 esi: c7cce800 edi: 00000007 ebp: c7cce800 esp: c7da5e70 (these vary) ds: 0018 es: 0018 ss: 0018 Process anaconda (pid: 11, stackpage=c7da5000) (always the same) Stack: 0000003f 000fba8b 00000260 c014c42b c3fbd0b0 00000007 c4e27a80 00003fe0 c36dac30 c3fbd0b0 c2032400 c765d7e0 0000003f c4e27a80 c72f7900 c7da5f08 5b513da2 c7da5f84 c611c020 c36dac30 fffffff3 c454d740 c454d740 c014dfe3 (stack contents vary as you might expect) Call Trace: [<co14c42b>] [<c014dfe3>] [c013901d>] [c0138eaa>] [<c01391aa>] [<c012e7f6>] [<c012eab9>] [<c0106dbf>] Code: 89 48 04 89 01 8b 46 58 8d 4e 58 89 50 04 89 43 08 89 4a 04 (both of those always the same)
Created attachment 914992 [details] Comment (This comment was longer than 65,535 characters and has been moved to an attachment by Red Hat Bugzilla).
Whoever os@gcx,de is, I'd appreciate it if next time, you avoided posting massive dumps of unrelated data to someone else's bug report... it looks to me like you have a different problem (yes it's a problem with anaconda but likely not the same one!). For example, I'm doing a network install, not a CD install, and from your logs there's clearly a problem with reading from your CD-ROM. Is there any way to split this into two bugs? thanks --Jim Morash
No unfortunately we can't remove the extra message. If you want to refile the bug we could do that, and close this one. BTW, have you tried a minimum install? The oops you are seeing looks like it may be running out of RAM, but with 128M that shouldn't be happening.
I agree that it seems unlikely that it's a RAM problem. I'm unable to experiment with the computer at the moment as it's in a robot that's in use (it boots fine with a pre-installed HD, but being able to install is an important capability). Should I just reopen this bug when I can do more testing?
Do you know if this machine's memory is fine? (You can test with something like memtest86)
I don't know. I may be able to test it sometime tomorrow. However, we've been running it on our robot for about eight hours a day for the last week and have had no crashes.
A quick comment on osa@cgx.de's bug, it is a failure of anaconda to ask for disk two, so anaconda crashes. It is bug #57842
Anaconda crashed on me (ouch, painful) during a similar situation. The differences are: 1 - I used the "Laptop" install selection 2 - I was installing over FTP I saved the crash data if that would be helpful.
If you are not the original bug poster please open a new bug with your issues. Putting back into NEEDINFO state because we are waiting on a response from the original bug poster.
Closing due to inactivity. Please reopen if you have more information.