1 |
Tom Stellard wrote: |
2 |
> On Mon, Mar 23, 2009 at 7:18 AM, Eric Thibodeau <kyron@××××××××.com> wrote: |
3 |
> |
4 |
>> In response to two e-mails I just got off list from Meetanshu Gupta and |
5 |
>> Ajith Kamath, in both cases the impression is that the project is simply to |
6 |
>> port the bootable CD to USB disk. Here is are some clarifications: |
7 |
>> |
8 |
>> - The wiki's¹ description is misleading, booting to USB was ONE of my |
9 |
>> suggestions with a little more details here². Quite frankly, porting to USB |
10 |
>> should be somewhat trivial. The hardest part is to being very methodical and |
11 |
>> making sure the process can be reproduced using, ideally, Catalyst and |
12 |
>> little outside scripting. All of this also implies getting the catalyst |
13 |
>> specs up to date, modifying/removing many hacks that would no longer be |
14 |
>> required for a USB based boot (since we would now have an RW system). *** |
15 |
>> root on JFFS2 might be of interest here ***... and so on. |
16 |
>> |
17 |
>> |
18 |
> |
19 |
> Is the idea that booting from a USB device would be added as an option |
20 |
> when building the liveCD image, or is booting from a USB device |
21 |
> supposed to replace booting from a CD? |
22 |
> |
23 |
The idea is to replace the CD. CDs are a PITA to use from an OS point of |
24 |
view and require overlayed over overlayed filesystems. Given the size |
25 |
price and availability of USB keys these days, it's just no longer worth |
26 |
the effort. Using a USB-based install also makes the installation |
27 |
"upgradable" and an excellent starting point that can evolve. We should |
28 |
then consider having the portage tree included in the image or make it |
29 |
easy to add it (ie: don't strip too much out of the image or have |
30 |
alternate images...which is a nuisance IMHO). |
31 |
> |
32 |
>> - AuFS is used as an overlay system and is required for the NFS-booted |
33 |
>> nodes. One might want to migrate that to a less efficient tmpfs approach |
34 |
>> (ref: LTSP implementation as being done under Gentoo too). |
35 |
>> - Change most of the scripts to use dialog and automate some detection/ask |
36 |
>> questions as per what was discovered. |
37 |
>> - ... and so on, the idea would be to implement my original proposal and |
38 |
>> push it further. I do have it but want some feedback before I dig it out and |
39 |
>> re-post it (Donnie, thought you had posted it as an example? If not so, feel |
40 |
>> free to do so if it's still relevant.) |
41 |
>> |
42 |
> |
43 |
> I have been working on building the liveCD the last few days, and I am |
44 |
> almost finished. I have run into many problems along the way, and I |
45 |
> have come up with a few general ideas for things to include in my |
46 |
> proposal: |
47 |
> |
48 |
> -Add spec files for x86. |
49 |
> |
50 |
yes |
51 |
> -Update the hpc-clustering and nfsroot profiles and the catalyst spec files. |
52 |
> |
53 |
yes |
54 |
> -Simplify and automate the build process with a configuration script. |
55 |
> |
56 |
? Meaning that you'd make a script that calls catalyst? Be my guest ;) |
57 |
> -Add documentation for building the liveCD using the new configuration script. |
58 |
> |
59 |
Or liveMedium as I started calling it for USB usage (or external HDD ;). |
60 |
Documentation is _always_ a good thing. |
61 |
> These tasks all involve the build process, and once they have been |
62 |
> completed, work could begin on USB booting, JFFS2 or other areas of |
63 |
> interest. |
64 |
> |
65 |
Booting off USB should be the priority, even before nice docs or |
66 |
automating the calling of catalyst. I didn't make a script for building |
67 |
the images since the script does very little once the environment is set up. |
68 |
> It seems like a lot of work has been done to make sure the right |
69 |
> packages are included in the build, and that currently the biggest |
70 |
> challenge is updating the build process and making it easy to |
71 |
> duplicate. |
72 |
I don't know the status of the new version of Catalyst but many new |
73 |
features, such as building stages 1-3 or even 1-4 in a single call, are |
74 |
supposed to be added. I suggest checking out the status of that project |
75 |
and then planning the "move" if applicable. Otherwise we'll have to |
76 |
stick to the hackerish scripts and catalyst specs. |
77 |
|
78 |
My notes on why duplicating the build process tend to point to the |
79 |
portage tree being so dynamic and the stage builds being very sensitive |
80 |
to the tree. With the advent of snapshots and stages being produced on a |
81 |
regular basis, the stage build pains should be lessened significantly |
82 |
compared to the quarterly or bi-annual releases. |
83 |
> Will improving the build process, USB booting and the |
84 |
> related file system changes be enough work for one summer, or are |
85 |
> there other parts of the cluster liveCD that Gentoo developers would |
86 |
> like to see improved? |
87 |
> |
88 |
Most of the configuration script(s) should be modified into a dialog(1) |
89 |
based script where the user is presented options that would be selected. |
90 |
The setup script I made is quite crude and reads in 2 config files. This |
91 |
should be changed to use dialog and provide the user the opportunity to |
92 |
change the default values. |
93 |
|
94 |
Another significant leap would be to be able to boot the nodes off |
95 |
NFSv4. This may sound trivial but the kernel deosn't support this |
96 |
natively which implies using a ramdisk (initrd). NFSv4 has many advantages: |
97 |
|
98 |
* Ability to use links outside of the exported root (server side): |
99 |
-- this means that most of the common config files between the |
100 |
server and nodes could be links...very _interesting_ |
101 |
* better caching |
102 |
* all sorts of new goodies I am forgetting about |
103 |
|
104 |
The problem with this might be that the overlayed filesystem used for |
105 |
the nfsroot might not like NFSv4 (aufs is already having a hard time |
106 |
with NFS at all). This might be yet another reason to use tmpfs+bind |
107 |
mounts*¹* as being done for the Gentoo MueCow project. |
108 |
|
109 |
> -Tom Stellard |
110 |
> |
111 |
So, there is still much to be done OS side even before we start looking |
112 |
at goodies such as integrating TAU² and other popular scientific |
113 |
apps...which mostly also imply good X drivers with all the surrounding |
114 |
issues... We might even have enough meat for more than one SoCer on this :P. |
115 |
|
116 |
Eric Thibodeau |
117 |
|
118 |
¹ http://www.funtoo.org/en/articles/linux/ffg/3/ "Tmpfs and bind mounts" |
119 |
² http://www.cs.uoregon.edu/research/tau/home.php |