It was not actually a loopback interface. But someone was using tcpdump utility without mentioning the snaplength and hence this error was received, which means pf_ring was trying to allocate ~409600*65535 which is almost 25 GB of memory. Its clear now!! Thanks a lot Alfredo for swift responses!
Regards, Gautam On Tue, Jan 9, 2018 at 2:48 PM, Alfredo Cardigliano <[email protected]> wrote: > That error was logged opening the loopback right? > > Alfredo > > > On 9 Jan 2018, at 10:13, Chandrika Gautam <[email protected]> > wrote: > > I will need one more information - > > As per code, slot len is calculated as below. > the_slot_len = pfr->slot_header_len + pfr->bucket_len; > = ALIGN(the_slot_len + sizeof(u_int16_t) /* RING_MAGIC_VALUE */, > sizeof(u_int64_t)); > > But below error logs slot length is logged as 65576, shouldn't this have > been quite smaller if MTU is set as 1500. > [PF_RING] ERROR: min_num_slots (409600, slot len = 65576) causes memory > size to wrap > > Regards, > Gautam > > On Tue, Jan 9, 2018 at 2:32 PM, Chandrika Gautam < > [email protected]> wrote: > >> Thanks Alfredo! >> >> Regards, >> Chandrika >> >> On Tue, Jan 9, 2018 at 2:13 PM, Alfredo Cardigliano <[email protected] >> > wrote: >> >>> Correct >>> >>> Alfredo >>> >>> >>> On 9 Jan 2018, at 08:38, Chandrika Gautam <[email protected]> >>> wrote: >>> >>> As per the code, it seems that all these memory allocation happens at >>> the time of ring creation only. >>> And a ring will be created whenever any new application(using pf_ring) >>> is spwaned. >>> And there should not be multiple occurrences of these errors if any >>> application is already up and running. >>> Please let me know if my understanding is correct. >>> >>> Regards, >>> Gautam >>> >>> On Tue, Jan 9, 2018 at 12:43 PM, Chandrika Gautam < >>> [email protected]> wrote: >>> >>>> Hi Alfredo, >>>> >>>> Does this error mean that packets are getting lost ? Problem is that I >>>> have not faced this issue in our lab. Its hit in one of the production >>>> server. >>>> Can you also explain what have you fixed to resolve this? >>>> >>>> Regards, >>>> Gautam >>>> >>>> On Sun, Jan 7, 2018 at 5:10 PM, Alfredo Cardigliano < >>>> [email protected]> wrote: >>>> >>>>> Hi Gautam >>>>> this is not related to hugepages (actually hugepages reserve physical >>>>> memory, thus >>>>> they can affect allocation, however in this specific case it was due >>>>> to a limit in the >>>>> pf_ring buffer size). >>>>> Please note I just pushed a patch to handle this case resizing the >>>>> buffer size when >>>>> limit is exceeded, please check if it’s working for you and let me >>>>> know. >>>>> >>>>> Regards >>>>> Alfredo >>>>> >>>>> >>>>> On 5 Jan 2018, at 06:38, Chandrika Gautam < >>>>> [email protected]> wrote: >>>>> >>>>> Hi Alfredo, >>>>> >>>>> Can this issue be faced if hugepages are not created on server? >>>>> Because these errors disappeared after creating the hugepages. >>>>> >>>>> Regards, >>>>> Gautam >>>>> >>>>> On Fri, Jan 5, 2018 at 10:52 AM, Chandrika Gautam < >>>>> [email protected]> wrote: >>>>> >>>>>> Hi Alfredo, >>>>>> >>>>>> MTU is 1518 for all 6 interfaces except for lo (loopback) >>>>>> interface(65535). >>>>>> >>>>>> Regards, >>>>>> Gautam >>>>>> >>>>>> >>>>>> On Tue, Jan 2, 2018 at 6:46 PM, Alfredo Cardigliano < >>>>>> [email protected]> wrote: >>>>>> >>>>>>> What is the MTU size? It seems you are trying to allocate more than >>>>>>> 25GB of memory, >>>>>>> thus you get this failure. Please try reducing the number of slots, >>>>>>> if you cannot reduce >>>>>>> the buffers size. >>>>>>> >>>>>>> Regards >>>>>>> Alfredo >>>>>>> >>>>>>> On 2 Jan 2018, at 10:40, Chandrika Gautam < >>>>>>> [email protected]> wrote: >>>>>>> >>>>>>> [PF_RING] ERROR: min_num_slots (409600, slot len = 65576) causes >>>>>>> memory size to wrap >>>>>>> [PF_RING] ring_mmap(): unable to allocate memory >>>>>>> >>>>>>> Regards, >>>>>>> Gautam >>>>>>> >>>>>>> On Tue, Jan 2, 2018 at 2:46 PM, Alfredo Cardigliano < >>>>>>> [email protected]> wrote: >>>>>>> >>>>>>>> Hi Gautam >>>>>>>> please provide the error, I see just the insmod command.. >>>>>>>> >>>>>>>> Regards >>>>>>>> Alfredo >>>>>>>> >>>>>>>> > On 27 Dec 2017, at 06:26, Chandrika Gautam < >>>>>>>> [email protected]> wrote: >>>>>>>> > >>>>>>>> > Hi, >>>>>>>> > >>>>>>>> > We are receiving below error while loading the pf_ring using >>>>>>>> below command even though there is free memory available on server. >>>>>>>> Please >>>>>>>> let me know if you need any other information. >>>>>>>> > >>>>>>>> > insmod pf_ring.ko min_num_slots=409600 enable_tx_capture=0 >>>>>>>> enable_frag_coherence=0 >>>>>>>> > >>>>>>>> > [root@CHOIPPROBE04B logs]# cat /proc/meminfo >>>>>>>> > MemTotal: 32743804 kB >>>>>>>> > MemFree: 24744428 kB >>>>>>>> > Buffers: 43208 kB >>>>>>>> > Cached: 445016 kB >>>>>>>> > SwapCached: 0 kB >>>>>>>> > Active: 6281204 kB >>>>>>>> > Inactive: 384396 kB >>>>>>>> > Active(anon): 6177672 kB >>>>>>>> > Inactive(anon): 8 kB >>>>>>>> > Active(file): 103532 kB >>>>>>>> > Inactive(file): 384388 kB >>>>>>>> > Unevictable: 0 kB >>>>>>>> > Mlocked: 0 kB >>>>>>>> > SwapTotal: 50331644 kB >>>>>>>> > SwapFree: 50331644 kB >>>>>>>> > Dirty: 2080 kB >>>>>>>> > Writeback: 0 kB >>>>>>>> > AnonPages: 6200104 kB >>>>>>>> > Mapped: 847312 kB >>>>>>>> > Shmem: 220 kB >>>>>>>> > Slab: 103712 kB >>>>>>>> > SReclaimable: 36832 kB >>>>>>>> > SUnreclaim: 66880 kB >>>>>>>> > KernelStack: 14208 kB >>>>>>>> > PageTables: 19664 kB >>>>>>>> > NFS_Unstable: 0 kB >>>>>>>> > Bounce: 0 kB >>>>>>>> > WritebackTmp: 0 kB >>>>>>>> > CommitLimit: 66703544 kB >>>>>>>> > Committed_AS: 9755140 kB >>>>>>>> > VmallocTotal: 34359738367 kB >>>>>>>> > VmallocUsed: 1437596 kB >>>>>>>> > VmallocChunk: 34340005064 kB >>>>>>>> > HardwareCorrupted: 0 kB >>>>>>>> > AnonHugePages: 6119424 kB >>>>>>>> > HugePages_Total: 0 >>>>>>>> > HugePages_Free: 0 >>>>>>>> > HugePages_Rsvd: 0 >>>>>>>> > HugePages_Surp: 0 >>>>>>>> > Hugepagesize: 2048 kB >>>>>>>> > DirectMap4k: 14336 kB >>>>>>>> > DirectMap2M: 2009088 kB >>>>>>>> > DirectMap1G: 31457280 kB >>>>>>>> > >>>>>>>> > Regards, >>>>>>>> > Gautam >>>>>>>> > _______________________________________________ >>>>>>>> > 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 >>>>>>> >>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> 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 >>>>> >>>> >>>> >>> _______________________________________________ >>> 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 > > > > _______________________________________________ > 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
