| 1 |
On Tue, Mar 23, 2021 at 10:23:11AM +0100, Michał Górny wrote: |
| 2 |
> On Sun, 2021-03-21 at 12:39 -0500, William Hubbs wrote: |
| 3 |
> > All, |
| 4 |
> > |
| 5 |
> > the following is a script which will migrate a Gentoo system to the usr |
| 6 |
> > merge layout. This is similar to the unsymlink-lib tool used to migrate |
| 7 |
> > a system from the 17.0 to the 17.1 profiles. |
| 8 |
> > |
| 9 |
> > I'm attaching it here to get some comments before I package it, so |
| 10 |
> > please let me know if I have missed something. |
| 11 |
> |
| 12 |
> To be honest, I don't think critical system modifications should be done |
| 13 |
> in shell script, and especially not via a fringe non-standards complaint |
| 14 |
> shell implementation in busybox. Even if you can assume you make no |
| 15 |
> mistakes, shell is unreliable by design. For example, your script may |
| 16 |
> start behaving in unexpected ways if you run out of space. |
| 17 |
|
| 18 |
I went with busybox shell in this case because busybox is a static |
| 19 |
binary and I figured with everything being moved around on the system it |
| 20 |
would be a good choice. |
| 21 |
|
| 22 |
I could pretty easily go with straight sh/bash, I just was focusing on |
| 23 |
bb since the commands are built in. |
| 24 |
|
| 25 |
You are right about shell behaving in weird ways if you run out of |
| 26 |
space, but that's true for anything that happens to be running when a |
| 27 |
system runs out of space. |
| 28 |
|
| 29 |
That is why I put the rollback directory in /var by |
| 30 |
default figuring there is more space there than in /, especially on |
| 31 |
systems with separate /var. |
| 32 |
|
| 33 |
The run_command wrapper in the script causes an immediate exit when the |
| 34 |
command it runs fails so things don't go any |
| 35 |
further than the failing command, so I'm not sure what else I can do |
| 36 |
with this situation. One option is to always echo the command we are |
| 37 |
about to run then just not run it if -d is specified. That means you |
| 38 |
would always see the commands the script is running. |
| 39 |
|
| 40 |
> You don't seem to be handling file collisions at all. Even today we |
| 41 |
> have files like /bin/bzip2 and /usr/bin/bzip2, not to mention shared |
| 42 |
> libraries. Silently ignoring the problem or requiring the users to |
| 43 |
> manually ensure their system is clean is not going to solve it. |
| 44 |
|
| 45 |
I have added -i to the cp commands in my latest version of this so I'll |
| 46 |
see when we have duplicate names, and yes that is an issue, even without |
| 47 |
the /usr merge. |
| 48 |
I'm curious why we are duplicating all of these names in the first |
| 49 |
place honestly. |
| 50 |
|
| 51 |
One example of this is in coreutils, and we even say in the ebuild |
| 52 |
comment that we need to figure out why we are doing it. |
| 53 |
|
| 54 |
There are a couple of ways forward for this. |
| 55 |
|
| 56 |
1) attempt to open bugs on the packages and remove the duplicate names |
| 57 |
by dropping symlinks and putting the files in their cannonical |
| 58 |
locations. this could lead to breakages if one of the alternate names is |
| 59 |
expected by something until the /usr merge is done. |
| 60 |
|
| 61 |
2) try to guess at resolving the duplicate names as part of the |
| 62 |
usr merge process. |
| 63 |
|
| 64 |
a. symmlinks in /bin or /sbin that point to the same name in /usr/bin or |
| 65 |
/usr/sbin should be removed. |
| 66 |
b. symlinks in /usr/bin or /usr/sbin that point to the same name in / |
| 67 |
should be removed. |
| 68 |
c. Other symlinks that I can think of should be preserved. |
| 69 |
|
| 70 |
Which path for handling this is best? Do you have any thoughts? |
| 71 |
|
| 72 |
> You don't seem to provide any helpful messages. When things fail, user |
| 73 |
> will be left in the blue with an error message from some system tool (or |
| 74 |
> rather, cheap-ass busybox rewrite, I guess). |
| 75 |
|
| 76 |
If the rollback directory is populated, you can use -r to recover, |
| 77 |
and the script will not let you perform the /usr merge if the rollback |
| 78 |
directory is not populated. |
| 79 |
|
| 80 |
> Also, have you verified that busybox's cp(1) actually preserves all file |
| 81 |
> properties (including xattrs, ACLs, caps...)? |
| 82 |
|
| 83 |
The documentation for busybox states that -p preserves file attributes |
| 84 |
if possible, and by doing ls -l in the chroot I can see that |
| 85 |
ownership/permissions are preserved. So, logically I would guess it |
| 86 |
preserves the others. |
| 87 |
|
| 88 |
If switching back to non-busybox is fine for this operation, I have no |
| 89 |
problem doing that either. |
| 90 |
|
| 91 |
> Please don't forget to include tests with it. Docker's good for testing |
| 92 |
> stuff like this. |
| 93 |
|
| 94 |
Docker can't be used during src_test that I'm aware of, so I'm not sure |
| 95 |
how Docker could be used to test this. |
| 96 |
|
| 97 |
William |
Archives