My Oracle Support Banner

DNS Server System Crashes (kernel panic) Regularly Running Late Solaris 11.3 (Doc ID 2529150.1)

Last updated on OCTOBER 12, 2019

Applies to:

Solaris Operating System - Version 11.3 to 11.4 [Release 11.0]
Information in this document applies to any platform.

Symptoms

A system operating as a DNS server repeatedly panics with a panic stack similar to this: 

==== panic thread: 0xffffa1c00a829680 ==== CPU: 48 ====
==== panic user (LWP_SYS) thread: 0xffffa1c00a829680  PID: 5063  on CPU: 48 ====
cmd: /opt/local/sbin/named -n 64
fmri: svc:/network/dns/server:default

<snip>

pc:      unix:vpanic_common+0x13a:  addq   $0xf0,%rsp

unix:vpanic_common+0x13a()
unix:0xfffffffffb8a5590()
unix:0xfffffffffb8567a3()
void unix:trap+0x3f1((struct regs *)0xfffffffc85bf36e0, (caddr_t)0, (processorid_t)0x30)
-- panic trap data  type: 0xd (General protection)
 addr                 0  rp   0xfffffffc85bf36e0
 trapno     0xd (General protection)
 err          0
 %rfl   0x10282 (negative|interrupt enable|resume)
 savfp 0xfffffffc85bf37f0
 savpc unix:atomic_add_32+0:  lock addl %esi,(%rdi)

 %rbp  0xfffffffc85bf37f0  %rsp  0xfffffffc85bf37d8
 %rip  unix:atomic_add_32+0:  lock addl %esi,(%rdi)

 0%rdi 0xdeadbeefdeadbeef  1%rsi                  1  2%rdx               0x10
 3%rcx                  1  4%r8  0xdeadbeefdeadbeef  5%r9                   1

 %rax          0xdeadbeef  %rbx  0xffffa1c029be0800
 %r10  0xdeadbeefdeadbeef  %r11                   0  %r12                   0
 %r13  0xffffa1c0146d3ac0  %r14                   1  %r15                   0

 %cs       0x30 (KDS_SEL)  %ds       0x4b (UCS_SEL)
 %es       0x4b (UCS_SEL)  %ss       0x38 (GDT_U32CODE,KPL)
 %fs          0 (KFS_SEL)  %gs          0 (KFS_SEL)
 fsbase 0xfffffc85bf37f0ff
 gsbase 0xfffffc85bf37f0ff
<trap>unix:atomic_add_32+0()
genunix:label_hold - frame recycled
void ip:ixa_safe_copy+0x88((ip_xmit_attr_t *)0xffffa1c029b8a200, (ip_xmit_attr_t *)0xffffa1c029be0800)
ip_xmit_attr_t *ip:conn_get_ixa_impl+0x6e((conn_t *)0xffffa1c0146d3ac0, (boolean_t)0, (int)1)
ip_xmit_attr_t *ip:conn_get_ixa+0x1a((conn_t *)0xffffa1c0146d3ac0, (boolean_t)0)
int ip:udp_send_cmn+0x118((sock_lower_handle_t)0xffffa1c0146d3ac0, (mblk_t *)0, (uio_t *)0xfffffffc85bf3e80, (struct msghdr *)0xfffffffc85bf3cd0, (cred_t *)0xffffa1c00a54d9c0)
int ip:udp_send_uio+0x34((sock_lower_handle_t)0xffffa1c0146d3ac0, (uio_t *)0xfffffffc85bf3e80, (struct msghdr *)0xfffffffc85bf3cd0, (cred_t *)0xffffa1c00a54d9c0)
int sockfs:so_sendmsg_cmn+0x1b8((struct sonode *)0xffffa1c0146b1000, (struct msghdr *)0xfffffffc85bf3cd0, (struct uio *)0xfffffffc85bf3e80, (struct cred *)0xffffa1c00a54d9c0, (int)0x8000, (boolean_t)1)
int sockfs:so_sendmsg+0xea((struct sonode *)0xffffa1c0146b1000, (struct msghdr *)0xfffffffc85bf3cd0, (struct uio *)0xfffffffc85bf3e80, (struct cred *)0xffffa1c00a54d9c0)
int sockfs:socket_sendmsg+0x64((struct sonode *)0xffffa1c0146b1000, (struct msghdr *)0xfffffffc85bf3cd0, (struct uio *)0xfffffffc85bf3e80, (cred_t *)0xffffa1c00a54d9c0)
ssize_t sockfs:sendit+0x1c0((int)0x245, (struct msghdr *)0xfffffffc85bf3cd0, (struct uio *)0xfffffffc85bf3e80, (int)0x8000)
ssize_t sockfs:sendmsg+0x294((int)0x245, (struct msghdr *)0xffff80ffbec59560, (int)0x8000)
unix:_syscall_invoke+0x30()
-- switch to user thread's user stack --

Notice the process sending a UDP datagram is named - a DNS server process.  

This issue may be triggered by any application sending UDP data, but DNS servers do so very often, to many destinations - we've seen more than one customer experience this issue when running a DNS server. 

Cause

To view full details, sign in with your My Oracle Support account.

Don't have a My Oracle Support account? Click to get started!


In this Document
Symptoms
Cause
Solution
References


My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.