Thank you for the suggestion. Will benchmark in my conditions. вт, 28 нояб. 2023 г. в 12:06, Fatih USTA <fatihusta86@xxxxxxxxx>: > > You can use type of list. Separate 64 elements per set. After that append into list. > > ipset create dummy_ifaces list:set > > ipset create dummy0_0 hash:net,iface > ipset create dummy0_1 hash:net,iface > ipset create dummy0_2 hash:net,iface > > ipset add dummy_ifaces dummy0_0 > ipset add dummy_ifaces dummy0_1 > ipset add dummy_ifaces dummy0_2 > > > On Tue, Nov 28, 2023, 12:34 Jozsef Kadlecsik <kadlec@xxxxxxxxxxxxx> wrote: >> >> On Tue, 28 Nov 2023, Марк Коренберг wrote: >> >> > Actually, I need an ipset that matches against list of interfaces >> > (without networks associated). Are there any ways ? >> >> No, that's not possible in ipset either. >> >> However, I'd suggest you to explore nftables where there are no such >> internal limitation than in ipset, supports matching interface indices or >> names and can store just interface names/indices in an nftables set too. >> >> Best regards, >> Jozsef >> > вт, 28 нояб. 2023 г. в 09:48, Jozsef Kadlecsik <kadlec@xxxxxxxxxxxxx>: >> > > >> > > Hi, >> > > >> > > On Tue, 28 Nov 2023, Марк Коренберг wrote: >> > > >> > > > for i in `seq 0 70`; do ip link del dummy$i; done; >> > > > for i in `seq 0 70`; do ip link add type dummy; done; >> > > > for i in `seq 0 70`; do ipset add qwe 0.0.0.0/0,dummy$i; done; >> > > > >> > > > Reveals the problem. Only 64 records can be added, but there are no >> > > > obvious restrictions on that. I s it possible to increase the limit ? >> > > >> > > It is intentional. Such elements can be stored in the same hash bucket >> > > only and 64 is the max size I'm willing to sacrifice for that. Please >> > > note, that's a huge number and means linear evaluation, i.e. loosing >> > > performance. >> > > >> > > Best regards, >> > > Jozsef >> > > -- >> > > E-mail : kadlec@xxxxxxxxxxxxxxxxx, kadlecsik.jozsef@xxxxxxxxx >> > > PGP key : https://wigner.hu/~kadlec/pgp_public_key.txt >> > > Address : Wigner Research Centre for Physics >> > > H-1525 Budapest 114, POB. 49, Hungary >> > >> > >> > >> > -- >> > Segmentation fault >> > >> >> -- >> E-mail : kadlec@xxxxxxxxxxxxxxxxx, kadlecsik.jozsef@xxxxxxxxx >> PGP key : https://wigner.hu/~kadlec/pgp_public_key.txt >> Address : Wigner Research Centre for Physics >> H-1525 Budapest 114, POB. 49, Hungary -- Segmentation fault
