Fingerprint record lost randomly

I found a way to log oom kills by watching dmesg while browsing a heavy web site using the Browser:

dmesg -w | egrep "lowmemorykiller|oom_reaper|^ {19}.*"

Using that I get output like this:

[140539.580703] lowmemorykiller: Killing 'ec_fdroid:tab21' (22126), adj 905,
                   to free 85860kB on behalf of 'HwBinder:3801_1' (23173) because
                   cache 475268kB is below limit 668640kB for oom_score_adj 0
                   Free memory is 647552kB above reserved.
                   Free CMA is 0kB
                   Total reserve is 136256kB
                   Total free pages is 71000kB
                   Total file cache is 553396kB
                   Slab Reclaimable is 123100kB
                   Slab UnReclaimable is 417072kB
                   Total Slab is 540172kB
                   GFP mask is 0x6280c2
                   Kernel Memory Reclaimable is 13556kB
                   Free Swap 699252kB
                   queue len is 1 of max 1 reason:0x2 margin:-16314
[140539.581611] lowmemorykiller: Killing 'AndroidUI' (22051), adj 900,
                   to free 148920kB on behalf of 'HwBinder:3801_1' (23173) because
                   cache 475268kB is below limit 668640kB for oom_score_adj 0
                   Free memory is 647552kB above reserved.
                   Free CMA is 0kB
                   Total reserve is 136256kB
                   Total free pages is 71000kB
                   Total file cache is 553396kB
                   Slab Reclaimable is 123100kB
                   Slab UnReclaimable is 417072kB
                   Total Slab is 540172kB
                   GFP mask is 0x6280c2
                   Kernel Memory Reclaimable is 13556kB
                   Free Swap 699252kB
                   queue len is 1 of max 1 reason:0x2 margin:-16314
[140539.587560] lowmemorykiller: Killing 'com.whatsapp' (23989), adj 104,
                   to free 78024kB on behalf of 'HwBinder:3801_1' (23173) because
                   cache 474812kB is below limit 668640kB for oom_score_adj 0
                   Free memory is 656144kB above reserved.
                   Free CMA is 0kB
                   Total reserve is 136256kB
                   Total free pages is 80104kB
                   Total file cache is 552940kB
                   Slab Reclaimable is 123100kB
                   Slab UnReclaimable is 417072kB
                   Total Slab is 540172kB
                   GFP mask is 0x6280c2
                   Kernel Memory Reclaimable is 13556kB
                   Free Swap 698740kB
                   queue len is 1 of max 1 reason:0x2 margin:-14038
[140539.591645] lowmemorykiller: Killing 'jolla-email' (31314), adj 103,
                   to free 137720kB on behalf of 'lipstick' (5980) because
                   cache 474132kB is below limit 668640kB for oom_score_adj 0
                   Free memory is 658960kB above reserved.
                   Free CMA is 0kB
                   Total reserve is 136256kB
                   Total free pages is 83632kB
                   Total file cache is 552256kB
                   Slab Reclaimable is 123100kB
                   Slab UnReclaimable is 417072kB
                   Total Slab is 540172kB
                   GFP mask is 0x6000c0
                   Kernel Memory Reclaimable is 13100kB
                   Free Swap 698484kB
                   queue len is 1 of max 1 reason:0x2 margin:-13156
[140539.594922] oom_reaper: reaped process 22126 (ec_fdroid:tab21), now anon-rss:0kB, file-rss:0kB, shmem-rss:24kB
[140539.623615] oom_reaper: reaped process 31314 (jolla-email), now anon-rss:0kB, file-rss:0kB, shmem-rss:84kB
[140539.646406] oom_reaper: reaped process 23989 (com.whatsapp), now anon-rss:0kB, file-rss:0kB, shmem-rss:2780kB
[140539.681563] oom_reaper: reaped process 22051 (AndroidUI), now anon-rss:0kB, file-rss:128kB, shmem-rss:760kB

The good thing is that I can now see every OOM kill. The bad thing is that it’s very frequent… It’s a few kills per minute, give or take.

Edit: This is pretty much offtopic, so I started a new thread instead.

4 Likes