Re: live-migration performance regression when using pmem

On Thu, Jun 12, 2025 at 03:34:35PM +, Chaney, Ben wrote: > On 5/14/25, 9:59 AM, "David Hildenbrand" > wrote: > > > > >Because we need to make sure the pmem kind memory data is synced > >after migration, we choose to call pmem_persist() when the migration > >finish.

Re: live-migration performance regression when using pmem

On 5/14/25, 9:59 AM, "David Hildenbrand" mailto:da...@redhat.com>> wrote: >Because we need to make sure the pmem kind memory data is synced >after migration, we choose to call pmem_persist() when the migration >finish. This will make sure the data of pmem is safe and will not >lose if power is o

Re: live-migration performance regression when using pmem

On Tue, May 13, 2025 at 03:48:06PM +, Chaney, Ben wrote: > On 5/12/25, 2:50 PM, "Peter Xu" > wrote: > > > > What you said makes sense to me, but I'm neither pmem user nor > > expert. Let's wait to see whether others would like to chime in. > > > > What's the first

Re: live-migration performance regression when using pmem

On 13.05.25 22:11, Michael S. Tsirkin wrote: On Tue, May 13, 2025 at 07:21:36PM +0200, David Hildenbrand wrote: On 12.05.25 17:16, Chaney, Ben wrote: Hello, When live migrating to a destination host with pmem there is a very long downtime where the guest is paused. In some cases, th

Re: live-migration performance regression when using pmem

On Tue, May 13, 2025 at 07:21:36PM +0200, David Hildenbrand wrote: > On 12.05.25 17:16, Chaney, Ben wrote: > > Hello, > > > > When live migrating to a destination host with pmem there is a > > very long downtime where the guest is paused. In some cases, this can be as > > high as 5 minu

Re: live-migration performance regression when using pmem

On 5/13/25, 1:21 PM, "David Hildenbrand" mailto:da...@redhat.com>> wrote: > Are you using a read-only NVDIMM? > In that case, I assume we would never need msync. We aren't using an NVDIMM at all, we are using pmem that is backed by DRAM. Ben

Re: live-migration performance regression when using pmem

On 12.05.25 17:16, Chaney, Ben wrote: Hello, When live migrating to a destination host with pmem there is a very long downtime where the guest is paused. In some cases, this can be as high as 5 minutes, compared to less than one second in the good case. Profiling suggests v

Re: live-migration performance regression when using pmem

On 5/12/25, 2:50 PM, "Peter Xu" mailto:pet...@redhat.com>> wrote: > What you said makes sense to me, but I'm neither pmem user nor > expert. Let's wait to see whether others would like to chime in. > What's the first bad commit of the regression? Is it since v10.0 release? Hi Peter, W

Re: live-migration performance regression when using pmem

On Mon, May 12, 2025 at 03:16:34PM +, Chaney, Ben wrote: > Hello, > > When live migrating to a destination host with pmem there is a very > long downtime where the guest is paused. In some cases, this can be as high > as 5 minutes, compared to less than one second in the good case. >

Re: live-migration performance regression when using pmem

On Mon, May 12, 2025 at 03:16:34PM +, Chaney, Ben wrote: > Hello, > > When live migrating to a destination host with pmem there is a very > long downtime where the guest is paused. In some cases, this can be as high > as 5 minutes, compared to less than one second in the good case. >

live-migration performance regression when using pmem

Hello, When live migrating to a destination host with pmem there is a very long downtime where the guest is paused. In some cases, this can be as high as 5 minutes, compared to less than one second in the good case. Profiling suggests very high activity in this code path: fff