From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jevon Qiao Subject: Re: when an osd is started up, IO will be blocked Date: Tue, 27 Oct 2015 23:26:44 +0800 Message-ID: <562F97B4.2090703@gmail.com> References: <5628F61E.1060708@gmail.com> <562DB9DB.1080200@gmail.com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1375857454==" Return-path: In-Reply-To: <562DB9DB.1080200-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: ceph-users-bounces-idqoXFIVOFJgJs9I8MT0rw@public.gmane.org Sender: "ceph-users" To: ceph-devel , ceph-users Cc: wangsongbo List-Id: ceph-devel.vger.kernel.org This is a multi-part message in MIME format. --===============1375857454== Content-Type: multipart/alternative; boundary="------------080708080500020507050802" This is a multi-part message in MIME format. --------------080708080500020507050802 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: quoted-printable Hi Cephers, We're in the middle of trying to triage the issue with ceph cluster=20 running 0.80.9 which was reported by Songbo and seeking for you experts'=20 advices. In fact, per our testing the process of stopping an working OSD and=20 starting it again will lead to a huge performance downgrade. In other=20 words, this issue can be reproduced quite easily, and we cannot lower=20 the impact of the state of OSD by tuning the settings like=20 osd_max_backfills/osd_recovery_max_chunk/osd_recovery_max_active.=20 Through looking into the source code, we notice that the requests issued=20 by clients will be queued firstly when the corresponding PGs are in some=20 certain states (like recovering and backfill) and then processed. During=20 this period, the IOPS outputted by fio drops significantly(from 2000 to=20 60). What we can think of this is to guarantee the data consistency, are=20 we correct? If that's the design, we're wondering how Ceph can support=20 the applications that are performance-sensitive? Is there any other=20 parameters we can tuning to reduce the impact? Thanks, Jevon On 26/10/15 13:27, wangsongbo wrote: > Hi all, > > When an osd is started, I will get a lot of slow requests from the=20 > corresponding osd log, as follows: > > 2015-10-26 03:42:51.593961 osd.4 [WRN] slow request 3.967808 seconds=20 > old, received at 2015-10-26 03:42:47.625968:=20 > osd_repop(client.2682003.0:2686048 43.fcf=20 > d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v=20 > 9744'347845) currently commit_sent > 2015-10-26 03:42:51.593964 osd.4 [WRN] slow request 3.964537 seconds=20 > old, received at 2015-10-26 03:42:47.629239:=20 > osd_repop(client.2682003.0:2686049 43.b4b=20 > cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v=20 > 9744'193029) currently commit_sent > 2015-10-26 03:42:52.594166 osd.4 [WRN] 40 slow requests, 17 included=20 > below; oldest blocked for > 53.692556 secs > 2015-10-26 03:42:52.594172 osd.4 [WRN] slow request 2.272928 seconds=20 > old, received at 2015-10-26 03:42:50.321151:=20 > osd_repop(client.3684690.0:191908 43.540=20 > f1858540/rbd_data.1fc5ca7429fc17.0000000000000280/head//43 v=20 > 9744'63645) currently commit_sent > 2015-10-26 03:42:52.594175 osd.4 [WRN] slow request 2.270618 seconds=20 > old, received at 2015-10-26 03:42:50.323461:=20 > osd_op(client.3684690.0:191911=20 > rbd_data.1fc5ca7429fc17.0000000000000209 [write 2633728~4096]=20 > 43.72b9f039 ack+ondisk+write e9744) currently commit_sent > 2015-10-26 03:42:52.594264 osd.4 [WRN] slow request 4.968252 seconds=20 > old, received at 2015-10-26 03:42:47.625828:=20 > osd_repop(client.2682003.0:2686047 43.b4b=20 > cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v=20 > 9744'193028) currently commit_sent > 2015-10-26 03:42:52.594266 osd.4 [WRN] slow request 4.968111 seconds=20 > old, received at 2015-10-26 03:42:47.625968:=20 > osd_repop(client.2682003.0:2686048 43.fcf=20 > d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v=20 > 9744'347845) currently commit_sent > 2015-10-26 03:42:52.594318 osd.4 [WRN] slow request 4.964841 seconds=20 > old, received at 2015-10-26 03:42:47.629239:=20 > osd_repop(client.2682003.0:2686049 43.b4b=20 > cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v=20 > 9744'193029) currently commit_sent > 2015-10-26 03:42:53.594527 osd.4 [WRN] 40 slow requests, 16 included=20 > below; oldest blocked for > 54.692945 secs > 2015-10-26 03:42:53.594533 osd.4 [WRN] slow request 16.004669 seconds=20 > old, received at 2015-10-26 03:42:37.589800:=20 > osd_repop(client.2682003.0:2686041 43.b4b=20 > cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v=20 > 9744'193024) currently commit_sent > 2015-10-26 03:42:53.594536 osd.4 [WRN] slow request 16.003889 seconds=20 > old, received at 2015-10-26 03:42:37.590580:=20 > osd_repop(client.2682003.0:2686040 43.fcf=20 > d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v=20 > 9744'347842) currently commit_sent > 2015-10-26 03:42:53.594538 osd.4 [WRN] slow request 16.000954 seconds=20 > old, received at 2015-10-26 03:42:37.593515:=20 > osd_repop(client.2682003.0:2686042 43.b4b=20 > cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v=20 > 9744'193025) currently commit_sent > 2015-10-26 03:42:53.594541 osd.4 [WRN] slow request 29.138828 seconds=20 > old, received at 2015-10-26 03:42:24.455641:=20 > osd_repop(client.4764855.0:65121 43.dbe=20 > 169a9dbe/rbd_data.49a7a4633ac0b1.0000000000000021/head//43 v=20 > 9744'12509) currently commit_sent > 2015-10-26 03:42:53.594543 osd.4 [WRN] slow request 15.998814 seconds=20 > old, received at 2015-10-26 03:42:37.595656:=20 > osd_repop(client.1800547.0:1205399 43.cc5=20 > 9285ecc5/rbd_data.1b794560c6e2ea.00000000000000d0/head//43 v=20 > 9744'36732) currently commit_sent > 2015-10-26 03:42:54.594892 osd.4 [WRN] 39 slow requests, 17 included=20 > below; oldest blocked for > 55.693227 secs > 2015-10-26 03:42:54.594908 osd.4 [WRN] slow request 4.273600 seconds=20 > old, received at 2015-10-26 03:42:50.321151:=20 > osd_repop(client.3684690.0:191908 43.540=20 > f1858540/rbd_data.1fc5ca7429fc17.0000000000000280/head//43 v=20 > 9744'63645) currently commit_sent > 2015-10-26 03:42:54.594911 osd.4 [WRN] slow request 4.271290 seconds=20 > old, received at 2015-10-26 03:42:50.323461:=20 > osd_op(client.3684690.0:191911=20 > rbd_data.1fc5ca7429fc17.0000000000000209 [write 2633728~4096]=20 > 43.72b9f039 ack+ondisk+write e9744) currently commit_sent > > Meanwhile, I run fio process with the rbd ioengine. > The iops of read and write were too small to response any IO from the=20 > fio process, > In other words, when an osd is started, the IO of the whole cluster=20 > will be blocked. > Is there some parameter to adjust ? > How to explain this problem? > The results of running fio process were as fllows: > > ebs: (g=3D0): rw=3Drandrw, bs=3D8K-8K/8K-8K/8K-8K, ioengine=3Drbd, iode= pth=3D64 > fio-2.2.9-20-g1520 > Starting 1 thread > rbd engine: RBD version: 0.1.9 > Jobs: 1 (f=3D1): [m(1)] [0.3% done] [0KB/0KB/0KB /s] [0/0/0 iops] [eta=20 > 05h:10m:14s] > ebs: (groupid=3D0, jobs=3D1): err=3D 0: pid=3D40323: Mon Oct 26 04:02:0= 0 2015 > read : io=3D10904KB, bw=3D175183B/s, *iops=3D21*, runt=3D 63737msec > slat (usec): min=3D0, max=3D61, avg=3D 1.11, stdev=3D 3.16 > clat (msec): min=3D1, max=3D63452, avg=3D1190.04, stdev=3D6046.28 > lat (msec): min=3D1, max=3D63452, avg=3D1190.04, stdev=3D6046.28 > clat percentiles (msec): > | 1.00th=3D[ 3], 5.00th=3D[ 4], 10.00th=3D[ 5], 20.00th= =3D[ =20 > 6], > | 30.00th=3D[ 6], 40.00th=3D[ 7], 50.00th=3D[ 8], 60.00th= =3D[ =20 > 9], > | 70.00th=3D[ 10], 80.00th=3D[ 12], 90.00th=3D[ 14], 95.00th= =3D[ 709], > | 99.00th=3D[16712], 99.50th=3D[16712], 99.90th=3D[16712],=20 > 99.95th=3D[16712], > | 99.99th=3D[16712] > bw (KB /s): min=3D 129, max=3D 2038, per=3D100.00%, avg=3D774.00,= =20 > stdev=3D1094.73 > write: io=3D24976KB, bw=3D401264B/s,*iops=3D48,* runt=3D 63737msec > slat (usec): min=3D0, max=3D40, avg=3D 2.48, stdev=3D 3.30 > clat (msec): min=3D2, max=3D31379, avg=3D786.91, stdev=3D4829.02 > lat (msec): min=3D2, max=3D31379, avg=3D786.92, stdev=3D4829.02 > clat percentiles (msec): > | 1.00th=3D[ 4], 5.00th=3D[ 6], 10.00th=3D[ 6], 20.00th= =3D[ =20 > 8], > | 30.00th=3D[ 9], 40.00th=3D[ 9], 50.00th=3D[ 11], 60.00th= =3D[ =20 > 12], > | 70.00th=3D[ 13], 80.00th=3D[ 15], 90.00th=3D[ 19], 95.00th= =3D[ =20 > 29], > | 99.00th=3D[16712], 99.50th=3D[16712], 99.90th=3D[16712],=20 > 99.95th=3D[16712], > | 99.99th=3D[16712] > bw (KB /s): min=3D 317, max=3D 5228, per=3D100.00%, avg=3D1957.33= ,=20 > stdev=3D2832.48 > lat (msec) : 2=3D0.07%, 4=3D3.23%, 10=3D52.29%, 20=3D35.76%, 50=3D4= .41% > lat (msec) : 750=3D1.36%, 1000=3D0.02%, 2000=3D0.02%, >=3D2000=3D2.= 83% > cpu : usr=3D0.03%, sys=3D0.00%, ctx=3D228, majf=3D0, minf=3D= 19 > IO depths : 1=3D0.1%, 2=3D0.1%, 4=3D0.4%, 8=3D3.9%, 16=3D18.9%, 32= =3D73.1%,=20 > >=3D64=3D3.5% > submit : 0=3D0.0%, 4=3D100.0%, 8=3D0.0%, 16=3D0.0%, 32=3D0.0%, = 64=3D0.0%,=20 > >=3D64=3D0.0% > complete : 0=3D0.0%, 4=3D97.5%, 8=3D0.0%, 16=3D0.2%, 32=3D0.9%, 6= 4=3D1.5%,=20 > >=3D64=3D0.0% > issued : total=3Dr=3D1363/w=3D3122/d=3D0, short=3Dr=3D0/w=3D0/d= =3D0,=20 > drop=3Dr=3D0/w=3D0/d=3D0 > latency : target=3D0, window=3D0, percentile=3D100.00%, depth=3D= 64 > > Run status group 0 (all jobs): > READ: io=3D10904KB, aggrb=3D171KB/s, minb=3D171KB/s, maxb=3D171KB/s,= =20 > mint=3D63737msec, maxt=3D63737msec > WRITE: io=3D24976KB, aggrb=3D391KB/s, minb=3D391KB/s, maxb=3D391KB/s,= =20 > mint=3D63737msec, maxt=3D63737msec > > Disk stats (read/write): > dm-0: ios=3D81/34, merge=3D0/0, ticks=3D472/26, in_queue=3D498,=20 > util=3D0.30%, aggrios=3D143/102, aggrmerge=3D106/122, aggrticks=3D1209/= 134,=20 > aggrin_queue=3D1343, aggrutil=3D0.60% > sdd: ios=3D143/102, merge=3D106/122, ticks=3D1209/134, in_queue=3D134= 3,=20 > util=3D0.60% > > > Thanks and Regards, > WangSongbo > > On 15/10/22 =E4=B8=8B=E5=8D=8810:43, wangsongbo wrote: >> Hi all, >> >> When an osd is started, relative IO will be blocked. >> According to the test result,the larger iops the clients send , the=20 >> longer it will take to elapse. >> Adjustment on all the parameters associate with recovery operations=20 >> was also found useless. >> >> How to reduce the impact of this process on the IO ? >> >> Thanks and Regards, >> WangSongbo >> > > --=20 > To unsubscribe from this list: send the line "unsubscribe ceph-devel" i= n > the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > More majordomo info at http://vger.kernel.org/majordomo-info.html --------------080708080500020507050802 Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable Hi Cephers,

We're in the middle of trying to triage the issue with ceph cluster running 0.80.9 which was reported by Songbo and seeking for you experts' advices.

In fact, per our testing the process of stopping an working OSD and starting it again will lead to a huge performance downgrade. In other words, this issue can be reproduced quite easily, and we cannot lower the impact of the state of OSD by tuning the settings like osd_max_backfills/osd_recovery_max_chunk/osd_recovery_max_active. Through looking into the source code, we notice that the requests issued by clients will be queued firstly when the corresponding PGs are in some certain states (like recovering and backfill) and then processed. During this period, the IOPS outputted by fio drops significantly(from 2000 to 60). What we can think of this is to guarantee the data consistency, are we correct? If that's the design, we're wondering how Ceph can support the applications that are performance-sensitive? Is there any other parameters we can tuning to reduce the impact?

Thanks,
Jevon
On 26/10/15 13:27, wangsongbo wrote:
Hi all,

When an osd is started, I will get a lot of slow requests from the corresponding osd log, as follows:

2015-10-26 03:42:51.593961 osd.4 [WRN] slow request 3.967808 seconds old, received at 2015-10-26 03:42:47.625968: osd_repop(client.2682003.0:2686048 43.fcf d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v 9744'347845) currently commit_sent
2015-10-26 03:42:51.593964 osd.4 [WRN] slow request 3.964537 seconds old, received at 2015-10-26 03:42:47.629239: osd_repop(client.2682003.0:2686049 43.b4b cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v 9744'193029) currently commit_sent
2015-10-26 03:42:52.594166 osd.4 [WRN] 40 slow requests, 17 included below; oldest blocked for > 53.692556 secs
2015-10-26 03:42:52.594172 osd.4 [WRN] slow request 2.272928 seconds old, received at 2015-10-26 03:42:50.321151: osd_repop(client.3684690.0:191908 43.540 f1858540/rbd_data.1fc5ca7429fc17.0000000000000280/head//43 v 9744'63645) currently commit_sent
2015-10-26 03:42:52.594175 osd.4 [WRN] slow request 2.270618 seconds old, received at 2015-10-26 03:42:50.323461: osd_op(client.3684690.0:191911 rbd_data.1fc5ca7429fc17.0000000000000209 [write 2633728~4096] 43.72b9f039 ack+ondisk+write e9744) currently commit_sent
2015-10-26 03:42:52.594264 osd.4 [WRN] slow request 4.968252 seconds old, received at 2015-10-26 03:42:47.625828: osd_repop(client.2682003.0:2686047 43.b4b cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v 9744'193028) currently commit_sent
2015-10-26 03:42:52.594266 osd.4 [WRN] slow request 4.968111 seconds old, received at 2015-10-26 03:42:47.625968: osd_repop(client.2682003.0:2686048 43.fcf d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v 9744'347845) currently commit_sent
2015-10-26 03:42:52.594318 osd.4 [WRN] slow request 4.964841 seconds old, received at 2015-10-26 03:42:47.629239: osd_repop(client.2682003.0:2686049 43.b4b cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v 9744'193029) currently commit_sent
2015-10-26 03:42:53.594527 osd.4 [WRN] 40 slow requests, 16 included below; oldest blocked for > 54.692945 secs
2015-10-26 03:42:53.594533 osd.4 [WRN] slow request 16.004669 seconds old, received at 2015-10-26 03:42:37.589800: osd_repop(client.2682003.0:2686041 43.b4b cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v 9744'193024) currently commit_sent
2015-10-26 03:42:53.594536 osd.4 [WRN] slow request 16.003889 seconds old, received at 2015-10-26 03:42:37.590580: osd_repop(client.2682003.0:2686040 43.fcf d1ddfcf/rbd_data.196483222ac2db.0000000000000010/head//43 v 9744'347842) currently commit_sent
2015-10-26 03:42:53.594538 osd.4 [WRN] slow request 16.000954 seconds old, received at 2015-10-26 03:42:37.593515: osd_repop(client.2682003.0:2686042 43.b4b cbcbbb4b/rbd_data.196483222ac2db.000000000000020b/head//43 v 9744'193025) currently commit_sent
2015-10-26 03:42:53.594541 osd.4 [WRN] slow request 29.138828 seconds old, received at 2015-10-26 03:42:24.455641: osd_repop(client.4764855.0:65121 43.dbe 169a9dbe/rbd_data.49a7a4633ac0b1.0000000000000021/head//43 v 9744'12509) currently commit_sent
2015-10-26 03:42:53.594543 osd.4 [WRN] slow request 15.998814 seconds old, received at 2015-10-26 03:42:37.595656: osd_repop(client.1800547.0:1205399 43.cc5 9285ecc5/rbd_data.1b794560c6e2ea.00000000000000d0/head//43 v 9744'36732) currently commit_sent
2015-10-26 03:42:54.594892 osd.4 [WRN] 39 slow requests, 17 included below; oldest blocked for > 55.693227 secs
2015-10-26 03:42:54.594908 osd.4 [WRN] slow request 4.273600 seconds old, received at 2015-10-26 03:42:50.321151: osd_repop(client.3684690.0:191908 43.540 f1858540/rbd_data.1fc5ca7429fc17.0000000000000280/head//43 v 9744'63645) currently commit_sent
2015-10-26 03:42:54.594911 osd.4 [WRN] slow request 4.271290 seconds old, received at 2015-10-26 03:42:50.323461: osd_op(client.3684690.0:191911 rbd_data.1fc5ca7429fc17.0000000000000209 [write 2633728~4096] 43.72b9f039 ack+ondisk+write e9744) currently commit_sent

Meanwhile, I run fio process with the rbd ioengine.
The iops of read and write were too small to response any IO from the fio process,
In other words, when an osd is started, the IO of the whole cluster will be blocked.
Is there some parameter to adjust ?
How to explain this=C2=A0 problem?
The results of running fio process were as fllows:

ebs: (g=3D0): rw=3Drandrw, bs=3D8K-8K/8K-8K/8K-8K, ioengine=3Drbd, iodepth=3D64
fio-2.2.9-20-g1520
Starting 1 thread
rbd engine: RBD version: 0.1.9
Jobs: 1 (f=3D1): [m(1)] [0.3% done] [0KB/0KB/0KB /s] [0/0/0 iops] [eta 05h:10m:14s]
ebs: (groupid=3D0, jobs=3D1): err=3D 0: pid=3D40323: Mon Oct 26 04:= 02:00 2015
=C2=A0 read : io=3D10904KB, bw=3D175183B/s, *iops=3D21*, runt=3D 63= 737msec
=C2=A0=C2=A0=C2=A0 slat (usec): min=3D0, max=3D61, avg=3D 1.11, std= ev=3D 3.16
=C2=A0=C2=A0=C2=A0 clat (msec): min=3D1, max=3D63452, avg=3D1190.04= , stdev=3D6046.28
=C2=A0=C2=A0=C2=A0=C2=A0 lat (msec): min=3D1, max=3D63452, avg=3D11= 90.04, stdev=3D6046.28
=C2=A0=C2=A0=C2=A0 clat percentiles (msec):
=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 1.00th=3D[=C2=A0=C2=A0=C2=A0 3],=C2= =A0 5.00th=3D[=C2=A0=C2=A0=C2=A0 4], 10.00th=3D[=C2=A0=C2=A0=C2=A0 5], 20.00th=3D[=C2=A0=C2=A0=C2=A0 6],
=C2=A0=C2=A0=C2=A0=C2=A0 | 30.00th=3D[=C2=A0=C2=A0=C2=A0 6], 40.00t= h=3D[=C2=A0=C2=A0=C2=A0 7], 50.00th=3D[=C2=A0=C2=A0=C2=A0 8], 60.00th=3D[=C2=A0=C2=A0=C2=A0 9],
=C2=A0=C2=A0=C2=A0=C2=A0 | 70.00th=3D[=C2=A0=C2=A0 10], 80.00th=3D[= =C2=A0=C2=A0 12], 90.00th=3D[=C2=A0=C2=A0 14], 95.00th=3D[ 709],
=C2=A0=C2=A0=C2=A0=C2=A0 | 99.00th=3D[16712], 99.50th=3D[16712], 99= .90th=3D[16712], 99.95th=3D[16712],
=C2=A0=C2=A0=C2=A0=C2=A0 | 99.99th=3D[16712]
=C2=A0=C2=A0=C2=A0 bw (KB=C2=A0 /s): min=3D=C2=A0 129, max=3D 2038,= per=3D100.00%, avg=3D774.00, stdev=3D1094.73
=C2=A0 write: io=3D24976KB, bw=3D401264B/s,*iops=3D48,* runt=3D 637= 37msec
=C2=A0=C2=A0=C2=A0 slat (usec): min=3D0, max=3D40, avg=3D 2.48, std= ev=3D 3.30
=C2=A0=C2=A0=C2=A0 clat (msec): min=3D2, max=3D31379, avg=3D786.91,= stdev=3D4829.02
=C2=A0=C2=A0=C2=A0=C2=A0 lat (msec): min=3D2, max=3D31379, avg=3D78= 6.92, stdev=3D4829.02
=C2=A0=C2=A0=C2=A0 clat percentiles (msec):
=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 1.00th=3D[=C2=A0=C2=A0=C2=A0 4],=C2= =A0 5.00th=3D[=C2=A0=C2=A0=C2=A0 6], 10.00th=3D[=C2=A0=C2=A0=C2=A0 6], 20.00th=3D[=C2=A0=C2=A0=C2=A0 8],
=C2=A0=C2=A0=C2=A0=C2=A0 | 30.00th=3D[=C2=A0=C2=A0=C2=A0 9], 40.00t= h=3D[=C2=A0=C2=A0=C2=A0 9], 50.00th=3D[=C2=A0=C2=A0 11], 60.00th=3D[=C2=A0=C2=A0 12],
=C2=A0=C2=A0=C2=A0=C2=A0 | 70.00th=3D[=C2=A0=C2=A0 13], 80.00th=3D[= =C2=A0=C2=A0 15], 90.00th=3D[=C2=A0=C2=A0 19], 95.00th=3D[=C2=A0=C2=A0 29],
=C2=A0=C2=A0=C2=A0=C2=A0 | 99.00th=3D[16712], 99.50th=3D[16712], 99= .90th=3D[16712], 99.95th=3D[16712],
=C2=A0=C2=A0=C2=A0=C2=A0 | 99.99th=3D[16712]
=C2=A0=C2=A0=C2=A0 bw (KB=C2=A0 /s): min=3D=C2=A0 317, max=3D 5228,= per=3D100.00%, avg=3D1957.33, stdev=3D2832.48
=C2=A0=C2=A0=C2=A0 lat (msec) : 2=3D0.07%, 4=3D3.23%, 10=3D52.29%, = 20=3D35.76%, 50=3D4.41%
=C2=A0=C2=A0=C2=A0 lat (msec) : 750=3D1.36%, 1000=3D0.02%, 2000=3D0= .02%, >=3D2000=3D2.83%
=C2=A0 cpu=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 : = usr=3D0.03%, sys=3D0.00%, ctx=3D228, majf=3D0, minf=3D19
=C2=A0 IO depths=C2=A0=C2=A0=C2=A0 : 1=3D0.1%, 2=3D0.1%, 4=3D0.4%, = 8=3D3.9%, 16=3D18.9%, 32=3D73.1%, >=3D64=3D3.5%
=C2=A0=C2=A0=C2=A0=C2=A0 submit=C2=A0=C2=A0=C2=A0 : 0=3D0.0%, 4=3D1= 00.0%, 8=3D0.0%, 16=3D0.0%, 32=3D0.0%, 64=3D0.0%, >=3D64=3D0.0%
=C2=A0=C2=A0=C2=A0=C2=A0 complete=C2=A0 : 0=3D0.0%, 4=3D97.5%, 8=3D= 0.0%, 16=3D0.2%, 32=3D0.9%, 64=3D1.5%, >=3D64=3D0.0%
=C2=A0=C2=A0=C2=A0=C2=A0 issued=C2=A0=C2=A0=C2=A0 : total=3Dr=3D136= 3/w=3D3122/d=3D0, short=3Dr=3D0/w=3D0/d=3D0, drop=3Dr=3D0/w=3D0/d=3D0
=C2=A0=C2=A0=C2=A0=C2=A0 latency=C2=A0=C2=A0 : target=3D0, window=3D= 0, percentile=3D100.00%, depth=3D64

Run status group 0 (all jobs):
=C2=A0=C2=A0 READ: io=3D10904KB, aggrb=3D171KB/s, minb=3D171KB/s, m= axb=3D171KB/s, mint=3D63737msec, maxt=3D63737msec
=C2=A0 WRITE: io=3D24976KB, aggrb=3D391KB/s, minb=3D391KB/s, maxb=3D= 391KB/s, mint=3D63737msec, maxt=3D63737msec

Disk stats (read/write):
=C2=A0=C2=A0=C2=A0 dm-0: ios=3D81/34, merge=3D0/0, ticks=3D472/26, = in_queue=3D498, util=3D0.30%, aggrios=3D143/102, aggrmerge=3D106/122, aggrticks=3D1209/134, aggrin_queue=3D1343, aggrutil=3D0.60%
=C2=A0 sdd: ios=3D143/102, merge=3D106/122, ticks=3D1209/134, in_qu= eue=3D1343, util=3D0.60%


Thanks and Regards,
WangSongbo

On 15/10/22 =E4=B8=8B=E5=8D=8810:43, wangsongbo wrote:
Hi all,

When an osd is started, relative IO will be blocked.
According to the test result,the larger iops the clients send , the longer it will take to elapse.
Adjustment on all the parameters associate with recovery operations was also found useless.

How to reduce the impact of this process on the IO ?

Thanks and Regards,
WangSongbo


--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at=C2=A0 http://vger.kernel.org/m= ajordomo-info.html

--------------080708080500020507050802-- --===============1375857454== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ ceph-users mailing list ceph-users-idqoXFIVOFJgJs9I8MT0rw@public.gmane.org http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com --===============1375857454==--