|
I've a Solaris box that act as a nat, but on a regular basis the system
crashes; some description follow: /etc/release: Solaris 9 12/03 s9s_u5wos_08b SPARC uname -a: SunOS host 5.9 Generic_122300-30 sun4u sparc SUNW,Sun-Fire-V210 scat panic: pc: ipf:nat_newmap+0x984: call ipf:nat_inlookup startpc: genunix:thread_create_intr+0x0: save %sp, -0xc0, %sp ipf:nat_newmap+0x984(0x2a100044cd0, 0x30002c0f148, 0x2a100044a00, 0xa00071a00000000, 0xc0a8006c00000000, 0x2a100044cd0) ipf:nat_new+0x2e0(0x2a100044cd0, 0x30002297c10, 0x0, 0x0, 0x1, 0x2a100044cd0) ipf:fr_checknatout+0x664(0x2a100044cd0, 0x2a100044cc4, 0x2a100044cd0, 0xfffff54a, 0x0, 0x0) ipf:fr_check+0x6f8(0x3000278a268, 0x14, 0x300011afd08, 0x1, 0x2a100044f70, 0x2a1000451b8) pfil:pfil_precheck+0xea8(0x30001174620, 0x2a1000451b8, 0x2, 0x300011afd08, 0xc0a8006c00000000, 0x3000278a284) pfil:pfilmodwput+0x26c(0x30001174620, 0x30003585d40, 0x20, 0x2a100045144, 0xf5c60800, 0x2298a10) unix:putnext+0x21c(0x300011adaa8?, 0x30003585d40, , 0x0, 0x8, 0x8) ip:ip_rput_forward+0x4bc(0x30003e5af50, 0x3000278a268, 0x30003585d40, 0x30000060cb0, 0x0, 0x8) ip:ip_rput+0x1558(0x30001175200?, 0x30003585d40, , , , 0x8) unix:putnext+0x21c(0x30001175490, 0x30003585d40, , 0x300011afed8, 0x16, 0x0) pfil:pfilmodrput+0x5d8(0x30001175490, 0x30003585d40, 0x20, 0x4, 0x16, 0x14) unix:putnext+0x21c(0x30001175720, 0x30003585d40, , 0x30000292cd6, 0xffffffffffffffff, 0x14) gld:gld_recv_tagged+0x204(0x30000d13800, 0x30003585d40, 0x0, 0x30001158000, 0x7, 0x22838ef1) gld:gld_recv() - frame recycled Reading it seems that a new allocation for a nat connection fails causing the kernel dump. The release of ipf is 4.1.33 (august 09) Some hints ? |

