[Mon Aug 31 15:31:55 2020] sysrq: Show Blocked State [Mon Aug 31 15:31:55 2020] task PC stack pid father [Mon Aug 31 15:31:55 2020] md1_reclaim D 0 806 2 0x80004000 [Mon Aug 31 15:31:55 2020] Call Trace: [Mon Aug 31 15:31:55 2020] __schedule+0x2dd/0x710 [Mon Aug 31 15:31:55 2020] ? finish_wait+0x80/0x80 [Mon Aug 31 15:31:55 2020] ? wbt_exit+0x30/0x30 [Mon Aug 31 15:31:55 2020] ? __wbt_done+0x30/0x30 [Mon Aug 31 15:31:55 2020] schedule+0x40/0xb0 [Mon Aug 31 15:31:55 2020] io_schedule+0x12/0x40 [Mon Aug 31 15:31:55 2020] rq_qos_wait+0xfa/0x170 [Mon Aug 31 15:31:55 2020] ? karma_partition+0x1e0/0x1e0 [Mon Aug 31 15:31:55 2020] ? wbt_exit+0x30/0x30 [Mon Aug 31 15:31:55 2020] wbt_wait+0x98/0xe0 [Mon Aug 31 15:31:55 2020] __rq_qos_throttle+0x23/0x30 [Mon Aug 31 15:31:55 2020] blk_mq_make_request+0x12a/0x5d0 [Mon Aug 31 15:31:55 2020] generic_make_request+0xcf/0x310 [Mon Aug 31 15:31:55 2020] submit_bio+0x42/0x1c0 [Mon Aug 31 15:31:55 2020] ? md_super_write.part.70+0x98/0x120 [md_mod] [Mon Aug 31 15:31:55 2020] md_update_sb.part.71+0x3c0/0x8f0 [md_mod] [Mon Aug 31 15:31:55 2020] r5l_do_reclaim+0x32a/0x3b0 [raid456] [Mon Aug 31 15:31:55 2020] ? schedule_timeout+0x162/0x340 [Mon Aug 31 15:31:55 2020] ? prepare_to_wait_event+0xb6/0x140 [Mon Aug 31 15:31:55 2020] ? md_register_thread+0xd0/0xd0 [md_mod] [Mon Aug 31 15:31:55 2020] md_thread+0x94/0x150 [md_mod] [Mon Aug 31 15:31:55 2020] ? finish_wait+0x80/0x80 [Mon Aug 31 15:31:55 2020] kthread+0x112/0x130 [Mon Aug 31 15:31:55 2020] ? kthread_park+0x80/0x80 [Mon Aug 31 15:31:55 2020] ret_from_fork+0x22/0x40 *shrug* These SSDs should be able to handle > 500MB/s. And > 130K IOPS. Swap would have to be pretty heavy to slow down journal writes. I'm not sure I have any good advise. My remaining ideas involve changing configuration just to see if the problem goes away, rather than actually understanding the cause of the problem. --- Chris Murphy
