* [PATCH] btrfs: When devices with same dev uuid occurs, only add the one with largest generation.
@ 2014-06-10 8:53 Qu Wenruo
2014-06-13 4:35 ` Anand Jain
0 siblings, 1 reply; 2+ messages in thread
From: Qu Wenruo @ 2014-06-10 8:53 UTC (permalink / raw)
To: linux-btrfs; +Cc: Anand Jain
Since btrfs currently use dulicated dev uuid when doing device replace,
the following problem will happen:
1) mount with device A missing using degraded mode.
2) replace device A with device B.
3) device A reappears.
4) umount btrfs fs.
5) mount btrfs fs.
After step 5), btrfs will still use device A even device B has a larger
generation.
The patch will judge generation when difference device with same dev
uuid.
And the patch should be applied after Anand's patch:
https://patchwork.kernel.org/patch/4309651/
Cc: Anand Jain <Anand.Jain@oracle.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
---
fs/btrfs/volumes.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index 0a5017a..07f0cf7 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -508,7 +508,8 @@ static noinline int device_list_add(const char *path,
ret = 1;
device->fs_devices = fs_devices;
- } else if (!device->name || strcmp(device->name->str, path)) {
+ } else if (!device->name || (strcmp(device->name->str, path) &&
+ found_transid > device->generation)) {
/*
* When FS is already mounted.
* 1. If you are here and if the device->name is NULL that means
--
2.0.0
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] btrfs: When devices with same dev uuid occurs, only add the one with largest generation.
2014-06-10 8:53 [PATCH] btrfs: When devices with same dev uuid occurs, only add the one with largest generation Qu Wenruo
@ 2014-06-13 4:35 ` Anand Jain
0 siblings, 0 replies; 2+ messages in thread
From: Anand Jain @ 2014-06-13 4:35 UTC (permalink / raw)
To: Qu Wenruo; +Cc: linux-btrfs
Qu,
Kindly note, device->generation is not updated when user scan the
device for the first time.
I have sent out patch:
[PATCH 2/2] btrfs: check generation as replace duplicates devid+uuid
Please feel free to add your Signed-off-by, if you think so.
review comments are appreciated as usual.
Thanks, Anand
On 10/06/2014 16:53, Qu Wenruo wrote:
> Since btrfs currently use dulicated dev uuid when doing device replace,
> the following problem will happen:
>
> 1) mount with device A missing using degraded mode.
> 2) replace device A with device B.
> 3) device A reappears.
> 4) umount btrfs fs.
> 5) mount btrfs fs.
>
> After step 5), btrfs will still use device A even device B has a larger
> generation.
>
> The patch will judge generation when difference device with same dev
> uuid.
> And the patch should be applied after Anand's patch:
> https://patchwork.kernel.org/patch/4309651/
>
> Cc: Anand Jain <Anand.Jain@oracle.com>
> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
> ---
> fs/btrfs/volumes.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
> index 0a5017a..07f0cf7 100644
> --- a/fs/btrfs/volumes.c
> +++ b/fs/btrfs/volumes.c
> @@ -508,7 +508,8 @@ static noinline int device_list_add(const char *path,
>
> ret = 1;
> device->fs_devices = fs_devices;
> - } else if (!device->name || strcmp(device->name->str, path)) {
> + } else if (!device->name || (strcmp(device->name->str, path) &&
> + found_transid > device->generation)) {
> /*
> * When FS is already mounted.
> * 1. If you are here and if the device->name is NULL that means
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-06-13 4:36 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-06-10 8:53 [PATCH] btrfs: When devices with same dev uuid occurs, only add the one with largest generation Qu Wenruo
2014-06-13 4:35 ` Anand Jain
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).