[Drbd-dev] [PATCH] drbd:clear NEW_CUR_UUID when uuid was actually created

[Drbd-dev] [PATCH] drbd:clear NEW_CUR_UUID when uuid was actually created

[Rui Xu]

If there is an IO blocked by creating new uuid, but we found
we dont have quorum, then NEW_CUR_UUID will be cleared but
new uuid was not actually created.

After one peer online, we have enough quorum to continue IO, then this
IO will reach  primary and the online peer.

But if there is another peer join later, this peer will found same uuid, and will lose data.

So don't clear NEW_CUR_UUID unless uuid was actually created.

Signed-off-by: Rui Xu <rui.xu@easystack.cn>
---
 drbd/drbd_sender.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/drbd/drbd_sender.c b/drbd/drbd_sender.c
index 7238730..b0e3701 100644
--- a/drbd/drbd_sender.c
+++ b/drbd/drbd_sender.c
@@ -2360,15 +2360,17 @@ void drbd_check_peers_new_current_uuid(struct drbd_device *device)
 
 	drbd_check_peers(resource);
 
-	if (device->have_quorum[NOW] && drbd_data_accessible(device, NOW))
+	if (device->have_quorum[NOW] && drbd_data_accessible(device, NOW)) {
 		drbd_uuid_new_current(device, false);
+		clear_bit(NEW_CUR_UUID, &device->flags);
+	}
 }
 
 static void make_new_current_uuid(struct drbd_device *device)
 {
 	drbd_check_peers_new_current_uuid(device);
 
-	get_work_bits(1UL << NEW_CUR_UUID | 1UL << WRITING_NEW_CUR_UUID, &device->flags);
+	clear_bit(WRITING_NEW_CUR_UUID, &device->flags);
 	wake_up(&device->misc_wait);
 }
 
-- 
1.8.3.1

Re: [Drbd-dev] [PATCH] drbd:clear NEW_CUR_UUID when uuid was actually created

[Joel Colledge]

Hi Xu,

You mentioned in your other message that this patch is intended for DRBD 9.0.
https://lists.linbit.com/pipermail/drbd-dev/2022-April/006577.html

I'm afraid that the 9.0 branch is end-of-life, so this won't be
applied. The drbd-9.1 branch is the place for improvements, especially
since it already contains related fixes to the handling of suspended
IO due to quorum loss.

Thank you for your participation in the project nevertheless!

Best regards,
Joel