[lxc-devel] lxc-clone outputs some logs
Serge Hallyn
serge.hallyn at ubuntu.com
Tue Nov 4 13:12:40 UTC 2014
Quoting KATOH Yasufumi (karma at jazz.email.ne.jp):
> >>> On Mon, 3 Nov 2014 15:44:32 +0000
> in message "Re: [lxc-devel] lxc-clone outputs some logs"
> Serge Hallyn-san wrote:
>
> > Quoting KATOH Yasufumi (karma at jazz.email.ne.jp):
> > > Hi,
> > >
> > > >>> On Thu, 30 Oct 2014 14:35:16 +0000
> > > in message "Re: [lxc-devel] lxc-clone outputs some logs"
> > > Serge Hallyn-san wrote:
> > >
> > > > > > > After applying commit edf7734 "overlay and aufs clone_paths: be more
> > > > > > > robust", clone was completed successfully but lxc-clone on unpriv env
> > > > > > > is now output the following log:
> > > > >
> > > > > > > $ lxc-clone -o ct02 -n test -s -B overlayfs
> > > > > > > lxc_container: conf.c: chown_mapped_root: 3649 Error stat /home/karma/.local/share/lxc/test/delta0
> > > > > > > Created container test as snapshot of ct02
> > > > > > > lxc_container: lxccontainer.c: copy_file: 2303 copy destination /home/karma/.local/share/lxc/test/lxc_rdepends exists
> > > > > > > Created container test as snapshot of ct02
> > > > >
> > > > > > ct02 is the cloned container with overlayfs.
> > > > >
> > > > > I find out other problem. After applying that patch, in part of the
> > > > > beginning of overlayfs_clonepath function, new->dest variable is,
> > > > > for example:
> > > > > "overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local/share/lxc/ct02/delta0"
> > > > >
> > > > > So, as the result, run "mkdir_p(new->dest, 0755)", and create dir in
> > > > > current dir.
> > > > >
> > > > > $ lxc-clone -o ct02 -n test -s -B overlayfs
> > > > > lxc_container: conf.c: chown_mapped_root: 3649 Error stat /home/karma/.local/share/lxc/test/delta0
> > > > > Created container test as snapshot of ct02
> > > > > lxc_container: lxccontainer.c: copy_file: 2303 copy destination /home/karma/.local/share/lxc/test/lxc_rdepends exists
> > > > > Created container test as snapshot of ct02
> > > > > $ find overlayfs\:/
> > > > > overlayfs:/
> > > > > overlayfs:/home
> > > > > overlayfs:/home/karma
> > > > > overlayfs:/home/karma/.local
> > > > > overlayfs:/home/karma/.local/share
> > > > > overlayfs:/home/karma/.local/share/lxc
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local/share
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local/share/lxc
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local/share/lxc/test
> > > > > overlayfs:/home/karma/.local/share/lxc/ct01/rootfs:/home/karma/.local/share/lxc/test/delta0
> > >
> > > > Sorry, I'm not really following, and I guess I haven't been using overlayfs
> > > > clones much lately. Do you how to fix it, and can you send a patch?
> > >
> > > This problem is also caused when clone by aufs. (between aufs/olfs)
> > >
> > > This problem is caused by passing invalid argument as bdev *new to
> > > (aufs|overlayfs)_clonepath.
> > >
> > > I don't understand your patch edf7734 well. But how about revert a
> > > part of the commit edf7734? as following:
> > >
>
> > Hm. I see. Ok, then we will have to fix the 'dir' case by actually
> > creating des if it doesn't exist. Else in the case of
>
> > lxc-clone -o x1 -n x2
>
> > where x1's config has
>
> > lxc.rootfs = /var/lib/lxc/x1/x
>
> > the clone will fail again, because dir_mount() tries to mount on a
> > nonexistant directory. Can you send the patch to fix that?
>
> Oh! OK, I understand.
>
> How about like this?
That looks good, thanks! Better than what I was thinking :)
Could you please resend on its own with a signed-off-by?
Also let's add a check to make sure that the strrchr doesn't
return NULL, because I don't see where we're ensuring that now.
It would crash somewhere anyway if that was the case, but
all the more reason to catch it.
>
> ---
> src/lxc/bdev.c | 21 ++++++++++++++++++++-
> 1 file changed, 20 insertions(+), 1 deletion(-)
>
> diff --git a/src/lxc/bdev.c b/src/lxc/bdev.c
> index ae5c77c..b303e00 100644
> --- a/src/lxc/bdev.c
> +++ b/src/lxc/bdev.c
> @@ -3328,6 +3328,7 @@ struct bdev *bdev_copy(struct lxc_container *c0, const char *cname,
> const char *oldname = c0->name;
> const char *oldpath = c0->config_path;
> struct rsync_data data;
> + char *rootfs;
>
> /* if the container name doesn't show up in the rootfs path, then
> * we don't know how to come up with a new name
> @@ -3338,12 +3339,30 @@ struct bdev *bdev_copy(struct lxc_container *c0, const char *cname,
> return NULL;
> }
>
> - orig = bdev_init(c0->lxc_conf, src, src, NULL);
> + orig = bdev_init(c0->lxc_conf, src, NULL, NULL);
> if (!orig) {
> ERROR("failed to detect blockdev type for %s", src);
> return NULL;
> }
>
> + if (!orig->dest) {
> + int ret;
> + orig->dest = malloc(MAXPATHLEN);
> + if (!orig->dest) {
> + ERROR("out of memory");
> + bdev_put(orig);
> + return NULL;
> + }
> + rootfs = strrchr(orig->src, '/');
> + rootfs++;
> + ret = snprintf(orig->dest, MAXPATHLEN, "%s/%s/%s", oldpath, oldname, rootfs);
> + if (ret < 0 || ret >= MAXPATHLEN) {
> + ERROR("rootfs path too long");
> + bdev_put(orig);
> + return NULL;
> + }
> + }
> +
> /*
> * special case for snapshot - if caller requested maybe_snapshot and
> * keepbdevtype and backing store is directory, then proceed with a copy
> --
> 2.1.1
> _______________________________________________
> lxc-devel mailing list
> lxc-devel at lists.linuxcontainers.org
> http://lists.linuxcontainers.org/listinfo/lxc-devel
More information about the lxc-devel
mailing list