On Tue 07-01-20 13:29:20, Chris Murphy wrote: > Hi, > > This is in response to: > https://lore.kernel.org/linux-fsdevel/20200104090955.GF23195@xxxxxxxxxxxxxxxxxxx/T/#m8b25fd42501d780d8053fc7aa9f4e3a28a19c49f > > I decided to open a bug report for tracking and attachments but I'm > also subscribed now to this list so - either here or there. > > "loss of responsiveness during heavy swap" > https://bugzilla.kernel.org/show_bug.cgi?id=206117 Please collect more snapshots of /proc/vmstat (e.g. in 1s intervals) since you start your workload. This will give us an insight on how the memory characteristics change over time. From the single snapshot we can see nr_zone_inactive_anon 247598 nr_zone_active_anon 1551129 nr_zone_inactive_file 25283 nr_zone_active_file 21563 pswpin 6002574 pswpout 11199783 which tells us that there is not all that much page cache resident (less than 200MB) while the majority of the memory is anonymous (7GB) most of it tracked as active. Btw. from a quick look at the sysrq output there seems to be quite a lot of tasks (more than 1k) running on the system. Only handful of them belong to the compilation. kswapd is busy and 13 processes in direct reclaim all swapping out to the disk. >From the above, my first guess would be that you are over subscribing memory you have available. I would focus on who is consuming all that memory. -- Michal Hocko SUSE Labs
