Hi Victor, Quoting lots having had a quick go at reformatting it.
> My old Acer Chromebook Model R11, circa December 2015, with a 1.6GHz quad-core Intel Celeron N3150 CPU. Is ChromeOS release R44 the latest update? > has got really, *really* slow. Like, every time I change tab it > re-loads the page. And If I am typing onto a page it lags several > seconds behind. Is that a really simple web page with just a HTML textarea, or one that's piled up with Javascript that runs on each keystroke? Is typing in the location bar also similarly slow? > At first I thought it was a memory issue, and unloaded various extensions > from Chrome but the thing continues to be slow even with a stripped down > Chrome, and very few tabs poen. > > chrome://system/ shows: > 1. crashes in dmesg log e.g. > [ 8757.474061] Task chrome:10706 refused to die (killer main:6598:6598, > nvcsw=2928, nivcsw=11253) > [ 8757.474084] chrome R running task 0 10706 5882 0x00100184 > [ 8757.474102] ffff880002aaf7f8 0000000000000086 0000000000000000 > ffff880014b2ee40 > [ 8757.474119] ffff88001b61d2b0 ffff88001b61d2b0 0000000000013b40 > ffff88001b61d2b0 > [ 8757.474136] ffffffffb636ff84 ffff880002aaffd8 ffff880029b50a50 > 0000000000000001 > [ 8757.474153] Call Trace: > [ 8757.474169] [<ffffffffb636ff84>] ? mmu_page_zap_pte+0xf/0x9f > [ 8757.474184] [<ffffffffb69e6e89>] preempt_schedule+0x36/0x51 > [ 8757.474198] [<ffffffffb65d3c76>] ___preempt_schedule+0x35/0x67 > [ 8757.474218] [<ffffffffb69eaadb>] ? _raw_spin_unlock+0x1f/0x21 > [ 8757.474230] [<ffffffffb637128a>] mmu_shrink_scan+0xd6/0x189 > [ 8757.474243] [<ffffffffb624638a>] shrink_slab+0x1dd/0x4ab > [ 8757.474257] [<ffffffffb649e7cc>] try_to_free_pages+0x295/0x59b > [ 8757.474272] [<ffffffffb623cb50>] __alloc_pages_nodemask+0x218c/0x2be6 > [ 8757.474286] [<ffffffffb6236544>] ? pagecache_get_page+0x89/0x470 > [ 8757.474300] [<ffffffffb625e029>] read_swap_cache_async+0x91/0x1e0 > [ 8757.474313] [<ffffffffb625e2dc>] swapin_readahead+0x164/0x187 > [ 8757.474326] [<ffffffffb6250915>] handle_mm_fault+0xfe8/0x206f > [ 8757.474340] [<ffffffffb6203481>] __do_page_fault+0x2ff/0x45a > [ 8757.474353] [<ffffffffb6200573>] ? __switch_to+0x158/0x367 > [ 8757.474365] [<ffffffffb6360e24>] ? kvm_on_user_return+0x53/0x55 > [ 8757.474378] [<ffffffffb62035e8>] do_page_fault+0xc/0xe > [ 8757.474389] [<ffffffffb69eceb2>] page_fault+0x22/0x30 > ... > [ 8890.707420] ThreadPoolForeg: page allocation failure: order:0, mode:0x200da > [ 8890.707446] CPU: 0 PID: 21782 Comm: ThreadPoolForeg Not tainted > 3.18.0-19748-gc785fae00b84 #1 > [ 8890.707459] Hardware name: GOOGLE Cyan, BIOS Google_Cyan.7287.57.209 > 06/13/2019 > [ 8890.707472] 0000000000000000 0000000062ad1553 ffff880028eff9c8 > ffffffffb69e5f68 > [ 8890.707490] 0000000000000000 00000000000200da ffff880028effa58 > ffffffffb649750d > [ 8890.707507] 0000004000000000 0000000000000246 ffff880028effa58 > ffffffffb64a3f01 > [ 8890.707524] Call Trace: > [ 8890.707541] [<ffffffffb69e5f68>] dump_stack+0x4e/0x71 > [ 8890.707555] [<ffffffffb649750d>] warn_alloc_failed+0x114/0x13d > [ 8890.707569] [<ffffffffb64a3f01>] ? wait_iff_congested+0xfb/0x165 > [ 8890.707583] [<ffffffffb621eda1>] ? __wake_up_sync+0x12/0x12 > [ 8890.707598] [<ffffffffb623cc83>] __alloc_pages_nodemask+0x22bf/0x2be6 > [ 8890.707613] [<ffffffffb6263d5f>] ? zs_unmap_object+0x15c/0x17f > [ 8890.707626] [<ffffffffb6236544>] ? pagecache_get_page+0x89/0x470 > [ 8890.707640] [<ffffffffb625e029>] read_swap_cache_async+0x91/0x1e0 > [ 8890.707653] [<ffffffffb625e2dc>] swapin_readahead+0x164/0x187 > [ 8890.707666] [<ffffffffb69eaaca>] ? _raw_spin_unlock+0xe/0x21 > [ 8890.707679] [<ffffffffb6250915>] handle_mm_fault+0xfe8/0x206f > [ 8890.707706] [<ffffffffb6203481>] __do_page_fault+0x2ff/0x45a > [ 8890.707719] [<ffffffffb6200623>] ? __switch_to+0x208/0x367 > [ 8890.707733] [<ffffffffb6360e24>] ? kvm_on_user_return+0x53/0x55 > [ 8890.707746] [<ffffffffb62035e8>] do_page_fault+0xc/0xe > [ 8890.707757] [<ffffffffb69eceb2>] page_fault+0x22/0x30 > ... > etc etc > > 2. meminfo looks ok... I agree; there's lot of Inactive pages that could be swapped out, and there's the swap space available. > MemTotal: 1992884 kB > MemFree: 225724 kB > MemAvailable: 464652 kB > Buffers: 84064 kB > Cached: 399772 kB > SwapCached: 131508 kB > Active: 787536 kB > Inactive: 358020 kB > Active(anon): 537752 kB > Inactive(anon): 179896 kB > Active(file): 249784 kB > Inactive(file): 178124 kB > Unevictable: 32436 kB > Mlocked: 34120 kB > SwapTotal: 4095996 kB > SwapFree: 3216904 kB > Dirty: 100 kB > Writeback: 0 kB > AnonPages: 682432 kB > Mapped: 252072 kB > Shmem: 55736 kB > Slab: 146864 kB > SReclaimable: 72564 kB > SUnreclaim: 74300 kB > KernelStack: 21264 kB > PageTables: 43176 kB > NFS_Unstable: 0 kB > Bounce: 0 kB > WritebackTmp: 0 kB > CommitLimit: 5092436 kB > Committed_AS: 30900496 kB > VmallocTotal: 34359738367 kB > VmallocUsed: 616564 kB > VmallocChunk: 34359117316 kB > > 3. memusage not an issue: > > Browser 229 MB, 493 FDs open of > 8192 > GPU [] 97 MB, 103 FDs open of > 8192 > Extension [Files] 76 MB, 42 FDs open of > 2048 > Extension [LastPass: Free Password Manager] 72 MB, 41 FDs open of > 2048 > Tab [chromebook /dev/root full - Google Search] 66 MB, 44 FDs open of > 2048 > Utility [Network Service] 60 MB, 52 FDs open of > 8192 > Extension [The Great Suspender] 53 MB, 45 FDs open of > 2048 > Extension [Trello] 44 MB, 39 FDs open of > 2048 > Extension [Stylus] 44 MB, 37 FDs open of > 2048 > Extension [Office Editing for Docs, Sheets & Slides] 39 MB, 30 FDs open of > 2048 > Extension [Text Mode] 37 MB, 36 FDs open of > 2048 > Tab (Chrome) [About System] 35 MB, 43 FDs open of > 2048 > Extension [WordWeb Dictionary Lookup] 35 MB, 30 FDs open of > 2048 > Utility [Assistant Audio Decoder Service] 28 MB, 29 FDs open of > 2048 > Pepper Plugin Broker [Shockwave Flash] 12 MB, 20 FDs open of > 8192 > Zygote 0 MB, 18 FDs open of > 2048 You've still got quite a few extensions running. > 4. But- aha! Disk is a problem -- but what is that file system? > > disk_usage > Filesystem 1K-blocks Used Available Use% Mounted on > /dev/root 1834352 1813120 21232 99% / This is a read-only filesystem that doesn't normally change thus its size is just big enough for the content. > devtmpfs 995308 0 995308 0% /dev > tmp 996440 168 996272 1% /tmp > run 996440 748 995692 1% /run > shmfs 996440 61652 934788 7% /dev/shm > /dev/mmcblk0p1 10801712 5520820 4712472 54% /mnt/stateful_partition This is where most of your data ends up. > /dev/mmcblk0p8 11760 24 11412 1% /usr/share/oem > > 5. swapinfo seems ok... > > swap_info > Filename Type Size Used Priority > /dev/zram0 partition 4095996 879972 -1 > low-memory margin (MiB): 101 778 > min_filelist_kbytes (KiB): 400000 > extra_free_kbytes (KiB): 0 > > edited highlights of system_files: > 2.1G /home/ > 1.9G /mnt/stateful_partition/ > > top memory use: high load average but low CPU > top memory > top - 13:29:10 up 15:56, 0 users, load average: 2.20, 5.75, 6.30 You've four cores, so presumably anything over four is CPU bound? > Tasks: 273 total, 1 running, 272 sleeping, 0 stopped, 0 zombie > %Cpu(s): 10.3 us, 12.8 sy, 0.0 ni, 76.9 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st 10.3 + 12.8 = 23.1% so almost one of the four cores is permanently busy, if it is doing all the work. And if it's a single-threaded thing burning CPU then it may be just one core. > MiB Mem : 1946.2 total, 425.0 free, 862.8 used, 658.4 buff/cache > MiB Swap: 4000.0 total, 3308.7 free, 691.3 used. 765.3 avail Mem ... > I am reluctant to install developer mode on the chromebook as that > will involve a powerwash factory reset and I have a few chrome > extensions I quite like, and stuff squirreled away in downloads that > I don't want to have to mess around with uploading to elsewhere. > > ANyone else have any idea on what I can do to remedy this ? Just > finding out what's happening and why would be a start. Given my normal machine for Firefox is a quad-core 1.8 GHz Intel Atom D525, I'd blame extra code running in the browser, e.g. Javascript for particular sits, and extensions. NoScript helped a lot here. :-) Does ChromeOS offer an about:about index of what there is to peruse, e.g. about:serviceworkers? Or performance monitoring by tab? I'd disable as much as easily possible to return to the Chromebook's more original state, knowing you can re-enable them easily too. See if that helps, and gradually bring things back. Web sites keep placing ever more load on the browser as developers often have the latest hardware, e.g. https://www.indiehackers.com/group/amas -- Cheers, Ralph. -- Next meeting: BEC, Bournemouth, Tuesday, 2019-12-03 20:00 Check to whom you are replying Meetings, mailing list, IRC, ... http://dorset.lug.org.uk/ New thread, don't hijack: mailto:dorset@mailman.lug.org.uk
