From: Logan Gunthorpe <logang@deltatee.com>
To: Robert LeBlanc <robert@leblancnet.us>
Cc: fio@vger.kernel.org, Jens Axboe <axboe@kernel.dk>
Subject: Re: fio: rdma_create_event_channel fail
Date: Mon, 4 Apr 2016 12:17:31 -0600 [thread overview]
Message-ID: <5702AFBB.5030800@deltatee.com> (raw)
In-Reply-To: <CAANLjFqdE=4zKV2J6JFehZHEXRgte-W8kVaT+5YAWD562kWcGA@mail.gmail.com>
Hey,
It looks like the ib verbs stuff isn't detecting your devices. Though
I'm not sure why. What you've sent so far looks good. Can you also send
the output of 'ibv_devices'? (It's from the ibverbs-utils package if you
don't have it installed.) You may have a missing or improperly installed
userspace driver and that command should inform you if that's the case.
It's still also worth trying some of the perftest tools (ib_write_bw) to
see if they have similar problems.
Logan
On 04/04/16 11:43 AM, Robert LeBlanc wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> Logan,
>
> Here is the output from your branch. Thanks for helping with this.
> Could the problem be due to us having multiple IB cards?
>
> # cat examples/rdmaio-server.fio
> # Example rdma server job
> [global]
> ioengine=rdma
> port=5557
> bs=1m
> size=100g
>
> [receiver]
> rw=read
> iodepth=16
>
> $ ./fio examples/rdmaio-server.fio
> receiver: (g=0): rw=read, bs=1M-1M/1M-1M/1M-1M, ioengine=rdma, iodepth=16
> fio-2.2.8-381-g2a44
> Starting 1 process
> fio: rdma_create_event_channel fail: No such device
> fio: io engine init failed. Perhaps try reducing io depth?
> fio: pid=5518, err=1/
>
>
> Run status group 0 (all jobs):
> [rleblanc@localhost fio-logan]$ ibstat
> CA 'mlx4_0'
> CA type: MT4099
> Number of ports: 1
> Firmware version: 2.35.5100
> Hardware version: 0
> Node GUID: 0x0cc47affff4fe8bc
> System image GUID: 0x0cc47affff4fe8bf
> Port 1:
> State: Active
> Physical state: LinkUp
> Rate: 56
> Base lid: 58
> LMC: 0
> SM lid: 1
> Capability mask: 0x02594868
> Port GUID: 0x0cc47affff4fe8bd
> Link layer: InfiniBand
> CA 'mlx5_0'
> CA type: MT4113
> Number of ports: 2
> Firmware version: 10.14.1100
> Hardware version: 0
> Node GUID: 0xe41d2d030006d0d0
> System image GUID: 0xe41d2d030006d0d0
> Port 1:
> State: Active
> Physical state: LinkUp
> Rate: 56
> Base lid: 56
> LMC: 0
> SM lid: 1
> Capability mask: 0x26596848
> Port GUID: 0xe41d2d030006d0d0
> Link layer: InfiniBand
> Port 2:
> State: Active
> Physical state: LinkUp
> Rate: 56
> Base lid: 57
> LMC: 0
> SM lid: 1
> Capability mask: 0x26596848
> Port GUID: 0xe41d2d030006d0d8
> Link layer: InfiniBand
>
> # lsmod
> Module Size Used by
> ebtable_filter 16384 0
> ebtables 36864 1 ebtable_filter
> ip6table_filter 16384 0
> ip6_tables 28672 1 ip6table_filter
> iptable_filter 16384 0
> iptable_raw 16384 0
> xprtrdma 53248 0
> ib_isert 57344 0
> iscsi_target_mod 294912 1 ib_isert
> ib_iser 53248 0
> ib_srpt 53248 0
> target_core_mod 372736 3 iscsi_target_mod,ib_srpt,ib_isert
> ib_srp 49152 0
> scsi_transport_srp 24576 1 ib_srp
> ib_ipoib 94208 0
> rdma_ucm 24576 0
> ib_ucm 24576 0
> ib_uverbs 49152 2 ib_ucm,rdma_ucm
> ib_umad 24576 0
> rdma_cm 45056 4 xprtrdma,ib_iser,rdma_ucm,ib_isert
> ib_cm 49152 5 rdma_cm,ib_srp,ib_ucm,ib_srpt,ib_ipoib
> iw_cxgb4 172032 0
> iw_cm 45056 2 iw_cxgb4,rdma_cm
> iw_cxgb3 126976 0
> mlx5_ib 110592 0
> mlx4_ib 151552 0
> ib_sa 36864 6 rdma_cm,ib_cm,mlx4_ib,ib_srp,rdma_ucm,ib_ipoib
> ib_mad 49152 5 ib_cm,ib_sa,mlx4_ib,ib_srpt,ib_umad
> ib_core 102400 19
> iw_cxgb3,iw_cxgb4,rdma_cm,ib_cm,ib_sa,iw_cm,xprtrdma,mlx4_ib,mlx5_ib,ib_mad,ib_srp,ib_ucm,ib_iser,ib_srpt,ib_umad,ib_uverbs,rdma_ucm,ib_ipoib,ib_isert
> ib_addr 20480 3 rdma_cm,ib_core,rdma_ucm
> ipmi_devintf 20480 2
> kvm_intel 155648 0
> kvm 495616 1 kvm_intel
> coretemp 16384 0
> iTCO_wdt 16384 0
> intel_powerclamp 16384 0
> iTCO_vendor_support 16384 1 iTCO_wdt
> sb_edac 28672 0
> x86_pkg_temp_thermal 16384 0
> joydev 20480 0
> sg 40960 0
> pcspkr 16384 0
> lpc_ich 24576 0
> edac_core 61440 1 sb_edac
> i2c_i801 20480 0
> mfd_core 16384 1 lpc_ich
> ipmi_si 57344 1
> 8250_fintek 16384 0
> acpi_power_meter 20480 0
> ipmi_msghandler 49152 2 ipmi_devintf,ipmi_si
> acpi_pad 180224 0
> mei_me 24576 0
> mei 90112 1 mei_me
> ioatdma 69632 0
> shpchp 40960 0
> ip_tables 28672 2 iptable_filter,iptable_raw
> xfs 929792 1
> libcrc32c 16384 1 xfs
> raid1 40960 1
> sd_mod 40960 2
> mlx4_en 110592 0
> vxlan 45056 1 mlx4_en
> ip6_udp_tunnel 16384 1 vxlan
> udp_tunnel 16384 1 vxlan
> crc32_pclmul 16384 0
> ast 61440 1
> syscopyarea 16384 1 ast
> sysfillrect 16384 1 ast
> aesni_intel 172032 0
> sysimgblt 16384 1 ast
> drm_kms_helper 126976 1 ast
> lrw 16384 1 aesni_intel
> gf128mul 16384 1 lrw
> glue_helper 16384 1 aesni_intel
> ablk_helper 16384 1 aesni_intel
> ttm 94208 1 ast
> igb 196608 0
> ahci 36864 0
> cryptd 20480 2 aesni_intel,ablk_helper
> ptp 20480 2 igb,mlx4_en
> crct10dif_pclmul 16384 0
> libahci 32768 1 ahci
> drm 352256 4 ast,ttm,drm_kms_helper
> crc32c_intel 24576 9
> mlx5_core 102400 1 mlx5_ib
> pps_core 20480 1 ptp
> mlx4_core 286720 2 mlx4_en,mlx4_ib
> libata 245760 2 ahci,libahci
> i2c_algo_bit 16384 2 ast,igb
> dca 16384 2 igb,ioatdma
> wmi 20480 0
> sunrpc 327680 2 xprtrdma
> dm_mirror 24576 0
> dm_region_hash 24576 1 dm_mirror
> dm_log 20480 2 dm_region_hash,dm_mirror
> dm_mod 110592 2 dm_log,dm_mirror
> iscsi_tcp 20480 4
> zfs 2826240 0
> be2iscsi 114688 0
> bnx2i 57344 0
> cnic 65536 1 bnx2i
> uio 20480 1 cnic
> cxgb4 204800 1 iw_cxgb4
> cxgb3 159744 1 iw_cxgb3
> libcxgbi 65536 0
> libiscsi_tcp 28672 2 iscsi_tcp,libcxgbi
> mdio 16384 1 cxgb3
> qla4xxx 286720 0
> iscsi_boot_sysfs 16384 2 qla4xxx,be2iscsi
> libiscsi 57344 7
> qla4xxx,libiscsi_tcp,bnx2i,be2iscsi,iscsi_tcp,ib_iser,libcxgbi
> scsi_transport_iscsi 102400 8
> qla4xxx,bnx2i,be2iscsi,iscsi_tcp,ib_iser,libcxgbi,libiscsi
> zunicode 331776 1 zfs
> zcommon 57344 1 zfs
> znvpair 94208 2 zfs,zcommon
> spl 98304 3 zfs,zcommon,znvpair
> zavl 16384 1 zfs
> -----BEGIN PGP SIGNATURE-----
> Version: Mailvelope v1.3.6
> Comment: https://www.mailvelope.com
>
> wsFcBAEBCAAQBQJXAqeiCRDmVDuy+mK58QAAeNoP/1pFkBV3piTGx9WPtXLA
> cK+SLIVirFcX1QcxJOSjXjYkTvHNwFgg5YzWbpQqN7ZzzEmt8baUiDeolTrD
> SwIKxE+SBS8qHGiFLndY2fO//poYevFAdq1JgYVGokv2eSvsrAJDR8snEJ/a
> qjA38VIIA3UoHE3ABDGgKO/vYE6dZZBfYHQRFIhp5tvDyFaCZzw8RGZWY7FR
> 9iTyQhONJxYV3oStTtmeM2B9txl+8HfdTTPpZXZsNZ0g8DKyF78gksBNfv4v
> HJYwLUhZBQNPlHhR82a9mxMvecaNBlyA33k4+uWjUieSX66YGwt+tn5vXl/k
> RY8QwchYI8v4642wPKCqXy+SAjJDn45wp4z/Z9Gx2cJfQFNw3c1rpHugINBF
> Ri+CS2IWG3ucCrn3K+Nqmu+SvH051j1xoyozZzBmLeMXILVLq2Fd93cjpN7r
> XVlmyZJ8kVyMoUwNkX/hRHVL4QRFP2vNKI2dOA0AfXgYOaMKrMUaAVmBWAMO
> XnyInJGR1ReUTZKMITdlNTn4lAoaYV3lug/1Uxk31T+hNhvS2uEBnK/JKv12
> QAO1Scl3dnD7uiMgtmoXp8jtCisd1N2jB6qblG4IVU22U8ROkfHxwmaHoHVr
> 5n4F/0WcY2SZVmwhKskX0bl0/bW8CWbeSb1LDEx67uPcTJk31Io9NE9rJQU6
> m4nq
> =pwG4
> -----END PGP SIGNATURE-----
> ----------------
> Robert LeBlanc
> PGP Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1
>
>
> On Mon, Apr 4, 2016 at 10:22 AM, Logan Gunthorpe <logang@deltatee.com> wrote:
>> Hey,
>>
>> I've created a patch on github that will print slightly more information
>> in the error message. If you could try it and post the output it may
>> shine some more light on the situation:
>>
>> https://github.com/axboe/fio/pull/160
>> https://github.com/lsgunth/fio/tree/rdma_err
>>
>> It may also be worth sending the output of ibstat and maybe lsmod as well.
>>
>> Logan
>>
>>
>> On 03/04/16 04:38 PM, Robert LeBlanc wrote:
>>> I'm able to do ibstat and everything is up. I can do ibping between the
>>> nodes and iSER is working properly. I have not done ib_write_bw but I
>>> can try it tomorrow if you think it would help.
>>>
>>> I included the entire message from fio. Can you give me some more
>>> information about additional error messages you may needed.
>>>
>>> Sent from a mobile device, please excuse any typos.
>>>
>>> On Apr 3, 2016 11:57 AM, "Logan Gunthorpe" <logang@deltatee.com
>>> <mailto:logang@deltatee.com>> wrote:
>>>
>>> Hi Robert,
>>>
>>> It looks like rdma_create_event_channel has failed which is a pretty
>>> basic part of the RDMA initialization. To me, this likely indicates
>>> your RDMA setup is broken. Do commands like ibstatus and ibv_devices
>>> report active interfaces? Do other RDMA test programs like
>>> ib_write_bw, etc (from the perftest package) work?
>>>
>>> It may be worth printing the errno with that error message, I may
>>> have time to make a patch to that effect tomorrow.
>>>
>>> Logan
>>>
>>> On 03/04/16 10:00 AM, Jens Axboe wrote:
>>>
>>> CC'ing Logan, who might have an idea.
>>>
>>>
>>> On 03/31/2016 10:44 AM, Robert LeBlanc wrote:
>>>
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA256
>>>
>>> When trying to use the RDMA engine and the example job for
>>> the server
>>> with the port set to an arbitrary value I get:
>>>
>>> # /home/rleblanc/fio/fio
>>> /home/rleblanc/fio/examples/rdmaio-server.fio
>>> receiver: (g=0): rw=read, bs=1M-1M/1M-1M/1M-1M,
>>> ioengine=rdma, iodepth=16
>>> fio-2.8-14-g23a8
>>> Starting 1 process
>>> fio: rdma_create_event_channel fail
>>> fio: io engine init failed. Perhaps try reducing io depth?
>>> fio: pid=18588, err=1/
>>>
>>> Setting I/O depth=1 only removed the corresponding message
>>> from the
>>> output.
>>> # uname -r
>>> 4.1.15.bs.ufd
>>>
>>> # rpm -qa | grep -E "ibverbs|rdmacm"
>>> libibverbs-devel-1.1.8-8.el7.x86_64
>>> librdmacm-1.0.21-1.el7.x86_64
>>> libibverbs-1.1.8-8.el7.x86_64
>>> librdmacm-devel-1.0.21-1.el7.x86_64
>>>
>>> I'm not sure what is wrong as this is the first time I've
>>> tried using
>>> the RDMA engine.
>>> - ----------------
>>> Robert LeBlanc
>>> PGP Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2
>>> FA62 B9F1
>>> -----BEGIN PGP SIGNATURE-----
>>> Version: Mailvelope v1.3.6
>>> Comment: https://www.mailvelope.com
>>>
>>> wsFcBAEBCAAQBQJW/VPPCRDmVDuy+mK58QAAPKMQAJF8B3y8pk9l2emtsrGm
>>> 2Rt7ufstv6c4XtuCk2wsc6ocZe8yNfAM1BNkW6pTF96orHZuLTt/QDvbDlnN
>>> q6N0vPkGJbDVbm7YNDzFc4qOU1pbrn8a66eck5BKuHPPogXCsJJTu+rdfAd9
>>> TNUGD4b9MzogTCzI8Zs6YRdWLIeaJRsPaHqJGYsD5G83rxGFagjx0qoOPuF+
>>> CNcFVYXZeU3+/YzsTDfuvNtiSDDJTUe3Shjw6fSu8ZFNabucAbbGOflovIIL
>>> kGjFmprrFgqOLiTnw7muF6tSXcc205YMGbCgOiEye4i9Ajd/ITiEQ3QlbQZ1
>>> WDz5WPSukDR8KqJoREKcksWVL7zVciulE5/+ZlJajD02JfOTz7j9QydLAPJ/
>>> sQM1g7Ft5HZK8TB9IgVKBernHCpahNQ5dU2OadDgpe0rxjzjrVcxegYOqLPd
>>> iUVFT2/UUFwzxaVnxHXTDNGO5A4JSyctvPTQ4uKLFFox9p6L5pFgrz9o86Fs
>>> lbW/72IjJD/8AEC64cqJp6JuC/sSEmz2hPpOvKdbpWPlVijzPB0OnpkMB4cC
>>> DATo4afT5uDRDe7IS8Ypi/WcriVLA+O9jRsigARri1F4FFc1QR/FDtXQnKYE
>>> SKB8sOE6sAv8qsyNLDqyD3rAzjRJ267zweNcWficcrHD3pYljXCLK1oaZgHD
>>> EcOB
>>> =f1X7
>>> -----END PGP SIGNATURE-----
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe
>>> fio" in
>>> the body of a message to majordomo@vger.kernel.org
>>> <mailto:majordomo@vger.kernel.org>
>>> More majordomo info at
>>> http://vger.kernel.org/majordomo-info.html
>>>
>>>
>>>
next prev parent reply other threads:[~2016-04-04 18:17 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-31 16:44 fio: rdma_create_event_channel fail Robert LeBlanc
2016-04-03 16:00 ` Jens Axboe
2016-04-03 17:57 ` Logan Gunthorpe
2016-04-03 22:38 ` Robert LeBlanc
2016-04-04 16:22 ` Logan Gunthorpe
2016-04-04 17:43 ` Robert LeBlanc
2016-04-04 18:17 ` Logan Gunthorpe [this message]
2016-04-04 18:48 ` Robert LeBlanc
2016-04-03 22:39 ` Robert LeBlanc
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5702AFBB.5030800@deltatee.com \
--to=logang@deltatee.com \
--cc=axboe@kernel.dk \
--cc=fio@vger.kernel.org \
--cc=robert@leblancnet.us \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.