Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-20 Thread monty
On Wed, Jun 20, 2018 at 10:51:17AM -0400, Mike Snitzer wrote: > > On Wed, Jun 20 2018 at 1:03pm -0400, > monty wrote: > > > > > On Tue, Jun 19, 2018 at 11:00:32AM -0400, Mike Snitzer wrote: > > > > > > On Tue, Jun 19 2018 at 10:43am -0400, > > > Joe Thornber wrote: > > > > > > > On Tue,

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-20 Thread Joe Thornber
On Wed, Jun 20, 2018 at 01:03:57PM -0400, monty wrote: > Hi, Mike and Joe. Thanks for your reply. I read __commit_transaction > many times and didn't find any problem of 2-phase commit. I use > md-raid1(PCIe nvme and md-raid5) in write-behind mode to store dm-thin > metadata. > Test case: > 1. I

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-20 Thread Mike Snitzer
On Wed, Jun 20 2018 at 1:03pm -0400, monty wrote: > > On Tue, Jun 19, 2018 at 11:00:32AM -0400, Mike Snitzer wrote: > > > > On Tue, Jun 19 2018 at 10:43am -0400, > > Joe Thornber wrote: > > > > > On Tue, Jun 19, 2018 at 09:11:06AM -0400, Mike Snitzer wrote: > > > > On Mon, May 21 2018 at

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-19 Thread monty
On Tue, Jun 19, 2018 at 11:00:32AM -0400, Mike Snitzer wrote: > > On Tue, Jun 19 2018 at 10:43am -0400, > Joe Thornber wrote: > > > On Tue, Jun 19, 2018 at 09:11:06AM -0400, Mike Snitzer wrote: > > > On Mon, May 21 2018 at 8:53pm -0400, > > > Monty Pavel wrote: > > > > > > > > > > > If

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-19 Thread Mike Snitzer
On Tue, Jun 19 2018 at 10:43am -0400, Joe Thornber wrote: > On Tue, Jun 19, 2018 at 09:11:06AM -0400, Mike Snitzer wrote: > > On Mon, May 21 2018 at 8:53pm -0400, > > Monty Pavel wrote: > > > > > > > > If dm_bufio_write_dirty_buffers func is called by __commit_transaction > > > func and

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-19 Thread Joe Thornber
On Tue, Jun 19, 2018 at 09:11:06AM -0400, Mike Snitzer wrote: > On Mon, May 21 2018 at 8:53pm -0400, > Monty Pavel wrote: > > > > > If dm_bufio_write_dirty_buffers func is called by __commit_transaction > > func and power loss happens during executing it, coincidencely > > superblock wrote

Re: [dm-devel] dm thin: superblock may write succeed before other metadata blocks because of wirting metadata in async mode.

2018-06-19 Thread Mike Snitzer
On Mon, May 21 2018 at 8:53pm -0400, Monty Pavel wrote: > > If dm_bufio_write_dirty_buffers func is called by __commit_transaction > func and power loss happens during executing it, coincidencely > superblock wrote correctly but some metadata blocks didn't. The reason > is we write all