Hi Alfredo, We are using SSD and wonder how can that be slow !! Please help to validate it!
Sent from my iPhone > On Aug 23, 2019, at 1:49 PM, Alfredo Cardigliano <[email protected]> wrote: > > Hi > please note the "WARNING: [sender] waiting.. (buffer empty)” messages in your > output, > this means that the IO is not fast enough reading data from disk and > providing it to disk2n, > what kind of storage are you using? > > Alfredo > >> On 23 Aug 2019, at 10:05, Chandrika Gautam <[email protected]> >> wrote: >> >> >> >> Sent from my iPhone >> >> Begin forwarded message: >> >>> From: Chandrika Gautam <[email protected]> >>> Date: August 23, 2019 at 1:34:05 PM GMT+5:30 >>> To: "[email protected]" <[email protected]> >>> Subject: Fw: Not able to achieve consistent 10Gbps on Napatech 2x40G card >>> for more than 10 seconds >>> >>> Hi, >>> >>> We are evaluating disk2n on Napatech 2x40G card to achieve sending traffic >>> more than 10G. >>> >>> We started for 10G traffic with 9 pcaps of around 1.5G to send on Napatech >>> port 0. >>> >>> Initially for few seconds we were able to observe 10G traffic then >>> afterwards it gradually reduces to 2.6Gbps and remains around 2-2.6Gbps for >>> rest of the period. >>> >>> So, we were not able to achieve consistent 10G traffic with disk2n on >>> Napatech card. >>> >>> We have also purchased the license for n2disk. >>> >>> Question. How can we achieve 10Gbps traffic on 2x40Napatech card with >>> driver ntanl_package_3gd-11.6.0-linux? >>> >>> OS details >>> >>> # uname -r >>> 3.10.0-957.1.3.el7.x86_64 >>> >>> # cat /etc/redhat-release >>> CentOS Linux release 7.6.1810 (Core) >>> >>> # numactl --hardware >>> available: 2 nodes (0-1) >>> node 0 cpus: 0 1 2 3 4 5 6 7 8 9 20 21 22 23 24 25 26 27 28 29 >>> node 0 size: 131034 MB >>> node 0 free: 127156 MB >>> node 1 cpus: 10 11 12 13 14 15 16 17 18 19 30 31 32 33 34 35 36 37 38 39 >>> node 1 size: 131071 MB >>> node 1 free: 125674 MB >>> node distances: >>> node 0 1 >>> 0: 10 20 >>> 1: 20 10 >>> >>> >>> >>> - O utput of disk2n is shared in attached file. >>> >>> >>> >>> root@RW-MUM-COUCHBASE1 ~]# >>> >>> We tried without caching option also (-b) and the max throughput was 2.5G >>> >>> # disk2n -i nt:0 -m /tmp/playpcap.txt -c 1 -w 2 -S 3 -C 1024 -I 450 -v >>> >>> While browsing the link >>> https://www.ntop.org/guides/pf_ring/modules/napatech.html >>> >>> in section >>> >>> 8.4. Napatech and Packet Copy >>> If you use the PF_RING (non-ZC) API packets are read in zero-copy. Instead >>> if you use PF_RING ZC API, a per-packet copy takes place, which is required >>> to move payload data from Napatech-memory to ZC memory. Keep this in mind! >>> >>> It describes that if we use PF_RING ZC then 1 copy happens. Could you >>> please explain this? >>> >>> -Iqbal >>> >> <observation.txt> >> _______________________________________________ >> Ntop-misc mailing list >> [email protected] >> http://listgateway.unipi.it/mailman/listinfo/ntop-misc > > _______________________________________________ > Ntop-misc mailing list > [email protected] > http://listgateway.unipi.it/mailman/listinfo/ntop-misc
_______________________________________________ Ntop-misc mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop-misc
