Date: Tue, 3 Jan 2017 17:21:21 +0100 From: Julien Cigar <julien@perdition.city> To: Meny Yossefi <menyy@mellanox.com> Cc: "freebsd-net@freebsd.org" <freebsd-net@freebsd.org>, Ben RUBSON <ben.rubson@gmail.com>, Yuval Bason <yuvalba@mellanox.com>, Hans Petter Selasky <hanss@mellanox.com> Subject: Re: FW: iSCSI failing, MLX rx_ring errors ? Message-ID: <20170103162120.GX15696@mordor.lan> In-Reply-To: <DB3PR05MB089011A41EF87A40C7AC741C36E0@DB3PR05MB089.eurprd05.prod.outlook.com> References: <486A6DA0-54C8-40DF-8437-F6E382DA01A8@gmail.com> <6a31ef00-5f7a-d36e-d5e6-0414e8b813c7@selasky.org> <DB3PR05MB089A5789A0A619FA8B7CA36C36C0@DB3PR05MB089.eurprd05.prod.outlook.com> <613AFD8E-72B2-4E3F-9C70-1D1E43109B8A@gmail.com> <2c9a9c2652a74d8eb4b34f5a32c7ad5c@AM5PR0502MB2916.eurprd05.prod.outlook.com> <DB3PR05MB089011A41EF87A40C7AC741C36E0@DB3PR05MB089.eurprd05.prod.outlook.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--6yuPXOSZRpyw7iEV Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable is it not the same issue as PR 211990 ? can you try by turning off jumbo frames ? On Tue, Jan 03, 2017 at 06:27:15AM +0000, Meny Yossefi wrote: >=20 > ________________________________________ > From: owner-freebsd-net@freebsd.orgOn Behalf OfBen RUBSON > Sent: Monday, January 2, 2017 11:09:15 AM (UTC+00:00) Monrovia, Reykjavik > To: freebsd-net@freebsd.org > Cc: Meny Yossefi; Yuval Bason; Hans Petter Selasky > Subject: Re: iSCSI failing, MLX rx_ring errors ? >=20 > Hi Meny, >=20 > Thank you very much for your feedback. >=20 > I think you are right, this could be a mbufs issue. > Here are some more numbers : >=20 > # vmstat -z | grep -v "0, 0$" > ITEM SIZE LIMIT USED FREE REQ FA= IL SLEEP > 4 Bucket: 32, 0, 2673, 28327, 88449799, 173= 17, 0 > 8 Bucket: 64, 0, 449, 15609, 13926386, 48= 71, 0 > 12 Bucket: 96, 0, 335, 5323, 10293892, 1428= 72, 0 > 16 Bucket: 128, 0, 533, 6070, 7618615, 4726= 47, 0 > 32 Bucket: 256, 0, 8317, 22133, 36020376, 5634= 79, 0 > 64 Bucket: 512, 0, 1238, 3298, 20138111, 114307= 42, 0 > 128 Bucket: 1024, 0, 1865, 2963, 21162182, 1587= 52, 0 > 256 Bucket: 2048, 0, 1626, 450, 80253784, 48901= 64, 0 > mbuf_jumbo_9k: 9216, 603712, 16400, 8744, 4128521064, 26= 61, 0 >=20 > # netstat -m > 32801/18814/51615 mbufs in use (current/cache/total) > 16400/9810/26210/4075058 mbuf clusters in use (current/cache/total/max) > 16400/9659 mbuf+clusters out of packet secondary zone in use (current/cac= he) > 0/8647/8647/2037529 4k (page size) jumbo clusters in use (current/cache/t= otal/max) > 16400/8744/25144/603712 9k jumbo clusters in use (current/cache/total/max) > 0/0/0/339588 16k jumbo clusters in use (current/cache/total/max) 188600K/= 137607K/326207K bytes allocated to network (current/cache/total) > 0/0/0 requests for mbufs denied (mbufs/clusters/mbuf+clusters) > 0/0/0 requests for mbufs delayed (mbufs/clusters/mbuf+clusters) > 0/0/0 requests for jumbo clusters delayed (4k/9k/16k) > 0/2661/0 requests for jumbo clusters denied (4k/9k/16k) > 0 sendfile syscalls > 0 sendfile syscalls completed without I/O request > 0 requests for I/O initiated by sendfile > 0 pages read by sendfile as part of a request > 0 pages were valid at time of a sendfile request > 0 pages were requested for read ahead by applications > 0 pages were read ahead by sendfile > 0 times sendfile encountered an already busy page > 0 requests for sfbufs denied > 0 requests for sfbufs delayed >=20 > I did not perform any mbufs tuning, numbers above are from FreeBSD itself. >=20 > This server has 64GB of memory. > It has a ZFS pool for which I limit ARC memory impact with : > vfs.zfs.arc_max=3D64424509440 #60G >=20 > The only thing I did is some TCP tuning to improve throughput over high-l= atency long-distance private links : > kern.ipc.maxsockbuf=3D7372800 > net.inet.tcp.sendbuf_max=3D6553600 > net.inet.tcp.recvbuf_max=3D6553600 > net.inet.tcp.sendspace=3D65536 > net.inet.tcp.recvspace=3D65536 > net.inet.tcp.sendbuf_inc=3D65536 > net.inet.tcp.recvbuf_inc=3D65536 > net.inet.tcp.cc.algorithm=3Dhtcp >=20 > Here are some graphs of memory & ARC usage when issue occurs. > Crosshair (vertical red line) is at the timestamp where I get iSCSI disco= nnections. > https://postimg.org/gallery/1kkekrc4e/ > What is strange is that each time issue occurs there is around 1GB of fre= e memory. > So FreeBSD should still be able to allocate some more mbufs ? > Unfortunately I do not have graphs about mbufs. >=20 > What should I ideally do ? >=20 > >> Have you tried increasing the mbufs limit?=20 > (sysctl) kern.ipc.nmbufs (Maximum number of mbufs allowed) >=20 >=20 > Thank you again, >=20 > Best regards, >=20 > Ben >=20 >=20 >=20 > > On 01 Jan 2017, at 09:16, Meny Yossefi <menyy@mellanox.com> wrote: > > > > Hi Ben, > > > > Those are not HW errors, note that: > > > > hw.mlxen1.stat.rx_dropped: 0 > > hw.mlxen1.stat.rx_errors: 0 > > > > It seems to be triggered when you are failing to allocate a replacement= buffer. > > Any chance you ran out of mbufs in the system? > > > > en_rx.c: > > > > mlx4_en_process_rx_cq(): > > > > mb =3D mlx4_en_rx_mb(priv, rx_desc, mb_list, length); > > if (!mb) { > > ring->errors++; > > goto next; > > } > > > > mlx4_en_rx_mb() =C3=A0 mlx4_en_complete_rx_desc(): > > > > /* Allocate a replacement page */ > > if (mlx4_en_alloc_buf(priv, rx_desc, mb_list, nr)) > > goto fail; > > > > -Meny > _______________________________________________ > freebsd-net@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-net > To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org" > _______________________________________________ > freebsd-net@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-net > To unsubscribe, send any mail to "freebsd-net-unsubscribe@freebsd.org" --=20 Julien Cigar Belgian Biodiversity Platform (http://www.biodiversity.be) PGP fingerprint: EEF9 F697 4B68 D275 7B11 6A25 B2BB 3710 A204 23C0 No trees were killed in the creation of this message. However, many electrons were terribly inconvenienced. --6yuPXOSZRpyw7iEV Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAABCgAdFiEE7vn2l0to0nV7EWolsrs3EKIEI8AFAlhrz30ACgkQsrs3EKIE I8CsiA/8D+q1LZJekx/l1E3Z2XuOp2xEst7DjgO0dInY4tMHDS32I5N9R7Ep5tXM PJNO/cm0kBrJtIXLTbQliN87MYk8HCd6ih6MGOkeaxSkcyj+dHQp3x1cd8i5orAq wiTvbyDXrcBPA5YYMqPeFxXbMEpoDf1iPGWzHk33vharGc7/TInEodic6/5c3B+s 2+HRatC0vubYmxt/HS5mRTomoKoHb9dc1CDhCB9SO2rlbGRsJ/QxGFFLssBJ3qxg vGo/Nc0k4sQ6zRG1uUbgYW/wfkeIEtqjiPdEeGCq+ydRnp8qNpV48N0vKfb048P+ TFwAu54TPs+J5Qt3P7Q9v/GU+SMBBgoYdns+TUbB5FirZSelg95LCZve8B3tE7d8 ZKyEbFS0Hihy5gt6a6+4E3Ht29ST326zFcXF+KWppwrlDGDOdh7FINH2nxW9SNJZ lxlSsw9iPe+F+oTz28IfJLOcDPszIEW2ZaQyQpRnGkbGmuic7C+QkUWdJGTqTSiV mc9iwv9d1MEcHfh6D8xVguV7eoiFVCe5a2lh7uQuvReK4ChtoFv1umIWG7lmx/IF ixDP4s2awSokuFVsCL8VX/ESiwsH2xiAZlWEgToWFxtQpANK7/4KH4dTYDjGOSOg JvWmjAinrEfpa2yY38yIo3pmI8WIzl6uoruZYgqzUtYfUPgMcIw= =e9ZB -----END PGP SIGNATURE----- --6yuPXOSZRpyw7iEV--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20170103162120.GX15696>