1 |
Yeah, thanks everyone for your input. I ran memtest86, and it appears |
2 |
that the last 30 megs of my ram fail one of the tests. Problem |
3 |
solved. I switched it with another stick I had lying around, and |
4 |
everything compiles just fine. Now to try and get a new stick under |
5 |
warranty.... |
6 |
|
7 |
As soon as you mentioned the idea that the ram might be bad, it seemed |
8 |
completely obvious that was the problem, even though I'd been puzzling |
9 |
over it for days. Thanks a lot people. |
10 |
|
11 |
|
12 |
On Fri, 26 Nov 2004 09:32:09 -0700, Duncan <1i5t5.duncan@×××.net> wrote: |
13 |
> Boyusmaximus posted <fec0e46d04112602396b094cb1@××××××××××.com>, excerpted |
14 |
> below, on Fri, 26 Nov 2004 10:39:10 +0000: |
15 |
> |
16 |
> |
17 |
> |
18 |
> > Hi, I'm having terrible problems reinstalling gentoo. Using the new |
19 |
> > 2004.3 minimal live cd, I progress to the point where I need to |
20 |
> > bootstrap my system, and I hit trouble. Everytime I try to run the |
21 |
> > bootstrap script, it fails at a random point along the way. I seem to |
22 |
> > always hit a segmentation fault or 'unnamed error' (my favourite kind) |
23 |
> > but at different points along the way, sometimes as early as building |
24 |
> > glibc, other times as late as building gcc. I guess this is some kind |
25 |
> > of hardware fault, |
26 |
> |
27 |
> Yes. Random points along the way almost certainly means hardware. |
28 |
> |
29 |
> What I did here is analyze the bootstrap script and execute the individual |
30 |
> steps one by one that it would have run. That was with 2004.0 (which I |
31 |
> never did get to work, because by the time I figured things out to the |
32 |
> point I wanted, 2004.1 was close enough that I decided to wait) and 2004.1 |
33 |
> (which /did/ work for me). Back then, and I imagine it's still the same, |
34 |
> the majority of the bootstrap script was simply checking what versions |
35 |
> were in the portage tree for it to build and that each one fit the |
36 |
> requirements of the profile. The actual bootstrap build process ended up |
37 |
> being about three lines of code (build one set with both use=bootstrap and |
38 |
> use=build set, unset one to build the next set, then do an environment |
39 |
> update necessary before building the third set) altho spread out in about |
40 |
> 15 lines of "glue" code near the end of the bootstrap script. |
41 |
> |
42 |
> Thus, what you do is |
43 |
> |
44 |
> 1) set all the environmental variables it would set, by hand (or manually |
45 |
> set an exit at an appropriate point in the bootstrap script, then "source" |
46 |
> it so it executes in the same command shell you are using, setting the |
47 |
> environment accordingly), ensuring build and bootstrap are in your USE |
48 |
> variable as well, then |
49 |
> |
50 |
> 2) emerge each package in turn of the first set |
51 |
> |
52 |
> 3) unset build or bootstrap (don't recall which, but its in the bootstrap |
53 |
> script so it's easy to see) |
54 |
> |
55 |
> 4) emerge each package in turn in the second set |
56 |
> |
57 |
> 5) update your environment using env-update, where the script would do so |
58 |
> |
59 |
> 6) continue, emerging each package in turn of the third set from the |
60 |
> bootstrap |
61 |
> |
62 |
> 7) tie up any config loose ends and unset the other one of |
63 |
> build/bootstrap, as the bootstrap script does to finish off. |
64 |
> |
65 |
> You are then left where you'd be if you'd just finished bootstrap, and can |
66 |
> resume the gentoo installation process from the handbook. |
67 |
> |
68 |
> Again, note that the above steps may have changed slightly since I did it |
69 |
> with 2004.1 -- the bootstrap script had changed a bit between that and |
70 |
> 2004.0, and that was before portage 2.0.51 and cascading profiles, so it's |
71 |
> likely changed a bit further now. However, it's certainly possible to |
72 |
> follow the bootstrap script for anyone that has a bit of bash scripting |
73 |
> under their belt, and the above demonstrates roughly what's involved, |
74 |
> besides actually parsing thru the script to understand what it's doing in |
75 |
> the first place. |
76 |
> |
77 |
> I'm actually very glad I ended up doing it that way myself, because I |
78 |
> ended up with a FAR better knowledge of the bootstrap process and the |
79 |
> infrastructure that makes Gentoo (and by extension, virtually all Linux |
80 |
> distributions) run, than I would have had if everything had "just worked" |
81 |
> as it was outlined in the manual. As well, I did this in my spare time |
82 |
> over several months (as I said, with several weeks off while waiting for |
83 |
> 2004.1), a single ebuild package at a time, all the while running from |
84 |
> Mandrake partitions on the same machine (which were eventually deleted |
85 |
> when I got Gentoo up and running to my satisfaction). Over all that extra |
86 |
> time, I was active in the various Gentoo groups (aka mailing lists, but I |
87 |
> participate thru GMane's list2news gateway so they are newsgroups to me), |
88 |
> including amd64 (naturally), user, and following announce and devel. |
89 |
> Therefore, by the time I got far enough to encounter most of the problems |
90 |
> other Gentoo newbies have (etc-updating away a site-specific fstab, |
91 |
> anyone?), I had seen them long before covered on one group or another, and |
92 |
> either avoided the issue or knew just what to do without myself having to |
93 |
> ask. In fact, as I was already a Mandrake power user, having been active |
94 |
> on their Cooker group/list for some time, and having migrated to amd64 on |
95 |
> Mandrake, I had the rather unusual experience of being able to answer a |
96 |
> number of questions in the user group in some detail, as well as |
97 |
> participate in the discussion here and a couple times in the devel group |
98 |
> too, well before I was even up and going on the Gentoo Linux platform! |
99 |
> |
100 |
> Thus, doing the bootstrap process manually instead of by script is not |
101 |
> only possible, but something I'd recommend any serious user intent on |
102 |
> learning about his system does at least once. It certainly benefited me! |
103 |
> |
104 |
> The caveat, of course, is that in ordered to do that, one must first have |
105 |
> a decent understanding of bash scripting. In my case, that wasn't an |
106 |
> issue, since I'd gained a working understanding of that early on in my |
107 |
> Mandrake experience, by dissecting their boot scripts, actually splitting |
108 |
> up sysinit.rc into multiple purpose based scripts as made sense to me, |
109 |
> then eventually doing an update to a later version of their script, and |
110 |
> similarly dissecting it, before giving up when I moved to Cooker and was |
111 |
> regularly updating, just as ~amd64 does on Gentoo. However, that |
112 |
> experience means I have at least a basic understanding of bash scripting, |
113 |
> enough to debug anyway, tho not necessarily enough to write my own scripts |
114 |
> of that complexity, and that was quite enough to take apart and |
115 |
> manually follow the Gentoo bootstrap script as well, which is exactly what |
116 |
> I did. |
117 |
> |
118 |
> Not everybody takes three months setting up their Gentoo system, but then |
119 |
> again, not everybody knows as much about it before they've even finished |
120 |
> stage three, as I did. |
121 |
> |
122 |
> -- |
123 |
> Duncan - List replies preferred. No HTML msgs. |
124 |
> "They that can give up essential liberty to obtain a little |
125 |
> temporary safety, deserve neither liberty nor safety." -- |
126 |
> Benjamin Franklin |
127 |
> |
128 |
> |
129 |
> |
130 |
> |
131 |
> -- |
132 |
> gentoo-amd64@g.o mailing list |
133 |
> |
134 |
> |
135 |
|
136 |
-- |
137 |
gentoo-amd64@g.o mailing list |