+bleung

Benson, I vaguely remember that you told me power manager would flush disk before going to sleep. Could you confirm it?

Thanks.

As mentioned in  bug 455749 , the system is syncing filesystem contents on suspend. Has this been proven to not be the case?

No. Just a suspicion based on comments from users. If we're sure the sync is happening, we're good.

I'd really like to see a test for this though, manual or otherwise.

powerd_suspend has explicit sync.

https://chromium.googlesource.com/chromiumos/platform2/+/master/power_manager/powerd/powerd_suspend

See,

log_msg "Explicit sync"
sync

A quick test on lumpy confirmed #2 and #4 that powerd would sync fs on suspecd:

2016-03-15T08:57:24.035126-07:00 NOTICE powerd_suspend[2760]: Going to suspend-to-RAM state: args=--suspend_duration=-1 --nosuspend_to_idle --wakeup_count=34676
2016-03-15T08:57:24.043065-07:00 INFO session_manager[857]: [INFO:policy_service.cc(188)] Persisted policy to disk.
2016-03-15T08:57:24.097416-07:00 INFO session_manager[857]: [INFO:child_exit_handler.cc(73)] 2776 is not a managed job.
2016-03-15T08:57:24.182345-07:00 NOTICE powerd_suspend[2896]: Lumpy with gobi modem
2016-03-15T08:57:24.185930-07:00 NOTICE powerd_suspend[2899]: Explicit sync
2016-03-15T08:57:27.084946-07:00 NOTICE powerd_suspend[2901]: Finalizing suspend
2016-03-15T08:58:36.129975-07:00 INFO kernel: [928646.706412] PM: Syncing filesystems ... done.
2016-03-15T08:58:36.129996-07:00 DEBUG kernel: [928646.758059] PM: Preparing system for mem sleep

I think we are good on suspend.