Hello,

I`am using openSuse 12.1 with kernel "Linux xenvm 3.1.7-default #1 SMP Sat Jan 7 17:12:53 CET 2012 x86_64 x86_64 x86_64 GNU/Linux"
and IPset 6.11 with 1GB RAM an 10 GB SWAP.
I`m trying to import all lot of IP ranges (more than 1.3 mio IP ranges found at iblocklist) .
First I create an IPset set with the command "ipset create antip2p hash:ip family inet hashsize 2000000 maxelem 4294967295"
when I try to set the IP ranges to ipset with the command "cat antip2p_list.txt | ipset restore" the following error apears

"ipset v6.11: Error in line 1: Kernel error received: Cannot allocate memory"

and in /var/log/messages:
May 3 17:58:45 xenvm kernel: [13191.708064] ipset: page allocation failure: order:0, mode:0x20
May 3 17:58:45 xenvm kernel: [13191.708072] Pid: 6241, comm: ipset Not tainted 3.1.7-default #1
May 3 17:58:45 xenvm kernel: [13191.708077] Call Trace:
May 3 17:58:45 xenvm kernel: [13191.708091] [<ffffffff81100ace>] warn_alloc_failed+0xee/0x160
May 3 17:58:45 xenvm kernel: [13191.708098] [<ffffffff811049c9>] __alloc_pages_nodemask+0x629/0x860
May 3 17:58:45 xenvm kernel: [13191.708106] [<ffffffff81141891>] kmem_getpages+0x51/0x150
May 3 17:58:45 xenvm kernel: [13191.708111] [<ffffffff81142293>] fallback_alloc+0x163/0x220
May 3 17:58:45 xenvm kernel: [13191.708117] [<ffffffff81141fc1>] ? cache_grow+0x2a1/0x2c0
May 3 17:58:45 xenvm kernel: [13191.708123] [<ffffffff81142072>] ____cache_alloc_node+0x92/0x150
May 3 17:58:45 xenvm kernel: [13191.708131] [<ffffffff81d07a07>] ? cache_alloc_refill+0x19e/0x1e5
May 3 17:58:45 xenvm kernel: [13191.708137] [<ffffffff8114294d>] __kmalloc+0x16d/0x1f0
May 3 17:58:45 xenvm kernel: [13191.708144] [<ffffffff81a4855f>] ? hash_ip4_elem_add.constprop.12+0x6f/0xc0
May 3 17:58:45 xenvm kernel: [13191.708151] [<ffffffff81a4855f>] hash_ip4_elem_add.constprop.12+0x6f/0xc0
May 3 17:58:45 xenvm kernel: [13191.708159] [<ffffffff81a48c59>] ? hash_ip4_resize+0x139/0x260
May 3 17:58:45 xenvm kernel: [13191.708164] [<ffffffff81a48c63>] hash_ip4_resize+0x143/0x260
May 3 17:58:45 xenvm kernel: [13191.708169] [<ffffffff81a46505>] ? hash_ip4_uadt+0x195/0x260
May 3 17:58:45 xenvm kernel: [13191.708176] [<ffffffff81a4108b>] call_ad+0xbb/0x260
May 3 17:58:45 xenvm kernel: [13191.708183] [<ffffffff81d28969>] ? _raw_spin_lock+0x9/0x10
May 3 17:58:45 xenvm kernel: [13191.708190] [<ffffffff81734020>] ? nla_parse+0x90/0xe0
May 3 17:58:45 xenvm kernel: [13191.708196] [<ffffffff81a41503>] ip_set_uadd+0x203/0x2b0
May 3 17:58:45 xenvm kernel: [13191.708203] [<ffffffff81b17fff>] ? in6_dump_addrs.isra.27+0xdf/0x190
May 3 17:58:45 xenvm kernel: [13191.708209] [<ffffffff81a142cf>] nfnetlink_rcv_msg+0x1ef/0x230
May 3 17:58:45 xenvm kernel: [13191.708215] [<ffffffff81a1410a>] ? nfnetlink_rcv_msg+0x2a/0x230
May 3 17:58:45 xenvm kernel: [13191.708220] [<ffffffff81a140e0>] ? nfnl_lock+0x20/0x20
May 3 17:58:45 xenvm kernel: [13191.708226] [<ffffffff81a10d69>] netlink_rcv_skb+0xa9/0xd0
May 3 17:58:45 xenvm kernel: [13191.708232] [<ffffffff819a2f1a>] ? __alloc_skb+0x4a/0x230
May 3 17:58:45 xenvm kernel: [13191.708237] [<ffffffff81a14010>] nfnetlink_rcv+0x10/0x20
May 3 17:58:45 xenvm kernel: [13191.708242] [<ffffffff81a10678>] netlink_unicast+0x2a8/0x2f0
May 3 17:58:45 xenvm kernel: [13191.708247] [<ffffffff819a2f4e>] ? __alloc_skb+0x7e/0x230
May 3 17:58:45 xenvm kernel: [13191.708253] [<ffffffff81a1098a>] netlink_sendmsg+0x2ca/0x360
May 3 17:58:45 xenvm kernel: [13191.708259] [<ffffffff81999836>] sock_sendmsg+0x106/0x120
May 3 17:58:45 xenvm kernel: [13191.708265] [<ffffffff819988be>] ? sock_destroy_inode+0x2e/0x40
May 3 17:58:45 xenvm kernel: [13191.708271] [<ffffffff8116ffc7>] ? destroy_inode+0x37/0x60
May 3 17:58:45 xenvm kernel: [13191.708277] [<ffffffff8199c5a9>] sys_sendto+0xf9/0x130
May 3 17:58:45 xenvm kernel: [13191.708284] [<ffffffff8117492b>] ? mntput_no_expire+0x2b/0xe0
May 3 17:58:45 xenvm kernel: [13191.708289] [<ffffffff811749fa>] ? mntput+0x1a/0x30
May 3 17:58:45 xenvm kernel: [13191.708295] [<ffffffff8115860e>] ? __fput+0x13e/0x200
May 3 17:58:45 xenvm kernel: [13191.708302] [<ffffffff81d30912>] system_call_fastpath+0x16/0x1b

IPset has about 16 mio. addresse stored in the set bevor the error occurs.
Is this a bug in IPset ? or is this to much addresses vor IPset ?
What is the maximum of addresses stored in one set ?

or is there an other solution to store millions of IP addresses/IP ranges ?
Can anybody please help me.

Best regards
B.-D.