Date: Fri, 19 Feb 2016 11:37:32 +0000 From: Steven Hartland <killing@multiplay.co.uk> To: freebsd-fs@freebsd.org Subject: Re: Zfs heavy io writing | zfskern txg_thread_enter Message-ID: <56C6FE7C.7080700@multiplay.co.uk> In-Reply-To: <CAM1TVW-yOvU6VM19PadD5ygsv2-Vb-_8T7SKjcsP7Ov0Q5A5SQ@mail.gmail.com> References: <CAM1TVW-yOvU6VM19PadD5ygsv2-Vb-_8T7SKjcsP7Ov0Q5A5SQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Please try 10.3-BETA2, FreeBSD 9.1 is so old its impossible to comment on its performance. On 19/02/2016 11:07, Niccolò Corvini wrote: > Hi, first time here! > We are having a problem with a server running FreeBsd 9.1 with ZFS on a > single sata drive. Since a few days ago, in the morning the system becomes > really slow due of a really heavy io writing. We investigated and we think > it might start at night, maybe correlated to to crondaily (standard) but we > are not sure. After a few hours the situation returns to normal. > Any help is much appreciated > The machine is a Intel Xeon E5-2620 with 36GB of RAM, the HDD is a 2TB an > is half full. > gstat output: > > L(q) ops/s r/s kBps ms/r w/s kBps ms/w %busy Name > 13 135 21 641 256.7 108 6410 41.4 128.8| ada0 > 0 0 0 0 0.0 0 0 0.0 0.0| ada0p1 > 13 135 21 641 256.7 108 6410 41.7 128.8| ada0p2 > 0 0 0 0 0.0 0 0 0.0 0.0| cd0 > 0 0 0 0 0.0 0 0 0.0 0.0| > gptid/3c0de011-4f37-11e5-8217-3085a91c3292 > 0 0 0 0 0.0 0 0 0.0 0.0| > zvol/zroot/swap > 13 135 21 641 256.7 108 6410 41.7 128.9| gpt/disk1 > > Using top -m io shows that the responsible is [zfskern{txg_thread_enter}] > top -m io output: > > PID JID USERNAME VCSW IVCSW READ WRITE FAULT TOTAL PERCENT COMMAND > 3 0 root 14 1 0 37 0 37 30.33% > [zfskern{txg_thread_enter}] > 49866 215 7070 26 2 0 5 0 5 4.10% > postgres: stats collector process (postgres) > 99901 5 70 42 0 0 4 0 4 3.28% > postgres: promeditec promeditec.osr.test 192.168.0.246(278 > 24820 199 www 10 0 7 0 0 7 5.74% > [jsvc{jsvc}] > 33869 212 88 19 2 0 2 0 2 1.64% > [mysqld{mysqld}] > 93400 0 root 13 0 10 0 0 10 8.20% [find] > 89407 215 7070 10 0 0 1 0 1 0.82% > postgres: alfresco alfconservazione.dotcom.ts.it 192.168.0 > 15776 5 70 11 0 0 4 0 4 3.28% > postgres: stats collector process (postgres) > 33869 212 88 10 0 0 3 0 3 2.46% > [mysqld{mysqld}] > 33869 212 88 2 0 0 11 0 11 9.02% > [mysqld{mysqld}] > 18685 198 root 5 0 0 2 0 2 1.64% > /usr/sbin/syslogd -s > 15852 214 70 4 1 0 1 0 1 0.82% > postgres: alfresco alfcomunets.dotcom.ts.it 192.168.0.212( > 98335 120 root 11 0 29 0 0 29 23.77% find > /var/log -name messages.* -mtime -2 > 16128 214 70 8 0 0 1 0 1 0.82% > postgres: alfresco alfaxErre8 192.168.0.208(50558) (postg > 1116 198 root 10 0 0 1 0 1 0.82% > sendmail: ./u1J9k90d001112 local: client DATA status (send > 1120 198 root 7 0 0 4 0 4 3.28% > mail.local -l > > Using procstat -kk on the zfskern pid shows: > > PID TID COMM TDNAME KSTACK > 3 100129 zfskern arc_reclaim_thre mi_switch sleepq_timedwait > _cv_timedwait arc_reclaim_thread fork_exit fork_trampoline > 3 100130 zfskern l2arc_feed_threa mi_switch sleepq_timedwait > _cv_timedwait l2arc_feed_thread fork_exit fork_trampoline > 3 100504 zfskern txg_thread_enter mi_switch sleepq_wait > _cv_wait txg_thread_wait txg_quiesce_thread fork_exit fork_trampoline > 3 100505 zfskern txg_thread_enter mi_switch sleepq_wait > _cv_wait zio_wait dsl_pool_sync spa_sync txg_sync_thread fork_exit > fork_trampoline > 3 100506 zfskern zvol zroot/swap mi_switch sleepq_wait _sleep > zvol_geom_worker fork_exit fork_trampoline > > systat -vmstat > > 7 users Load 0.50 0.62 1.46 Feb 19 11:46 > > Mem:KB REAL VIRTUAL VN PAGER SWAP > PAGER > Tot Share Tot Share Free in out in > out > Act 41318k 199492 1251183k 588716 2254748 count > All 46844k 229048 -1968M 892088 pages > Proc: Interrupts > r p d s w Csw Trp Sys Int Sof Flt cow 2836 total > 4k 8516 393 12k 236 1445 11k 11281 zfod > atkbd0 1 > ozfod acpi0 > 9 > 4.4%Sys 0.0%Intr 0.3%User 0.0%Nice 95.4%Idle %ozfod ehci0 > 17 > | | | | | | | | | | | daefr ehci1 > 23 > == 11171 prcfr 79 > cpu0:timer > dtbuf 11265 totfr isci0 > 264 > Namei Name-cache Dir-cache 1095774 desvn react 24 > em0:rx 0 > Calls hits % hits % 409282 numvn pdwak 16 > em0:tx 0 > 78 41 53 273943 frevn pdpgs > em0:link > intrn 196 ahci0 > 278 > Disks ada0 cd0 pass0 pass1 20445132 wire 267 > cpu21:time > KB/t 2.55 0.00 0.00 0.00 37317552 act 55 > cpu13:time > tps 223 0 0 0 4948708 inact 86 > cpu5:timer > MB/s 0.56 0.00 0.00 0.00 884804 cache 24 > cpu12:time > %busy 94 0 0 0 1370012 free 63 > cpu10:time > buf 306 > cpu19:time > 63 > cpu11:time > 55 > cpu14:time > 86 > cpu9:timer > 71 > cpu18:time > 86 > cpu3:timer > 47 > cpu23:time > 55 > cpu6:timer > 55 > cpu22:time > 71 > cpu2:timer > 39 > cpu20:time > > zpool status: > pool: zroot > state: ONLINE > scan: scrub repaired 0 in 3h46m with 0 errors on Wed Nov 4 21:54:44 2015 > config: > > NAME STATE READ WRITE CKSUM > zroot ONLINE 0 0 0 > gpt/disk1 ONLINE 0 0 0 > > errors: No known data errors > _______________________________________________ > freebsd-fs@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-fs > To unsubscribe, send any mail to "freebsd-fs-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56C6FE7C.7080700>