It seemed I was getting close to a solution on this issue, so I'm posting this
again in the hope that someone will know what's going wrong.
This IPv6 ping doesn't work, and it's causing a lengthy freeze in the tomcat
start-up (see bottom).
This is on Debian Etch linux 2.6.20 - can anyone
Admittedly grasping at straws here, but does an IPv6 ping to localhost work
the same on both boxes, e.g.:
[EMAIL PROTECTED]:~$ ping6 ::1
PING ::1(::1) 56 data bytes
64 bytes from ::1: icmp_seq=1 ttl=64 time=0.047 ms
64 bytes from ::1: icmp_seq=2 ttl=64 time=0.039 ms
--- ::1 ping statistics ---
2
No, on the box that doesn't work, I get this:
[EMAIL PROTECTED]:~$ sudo ping6 ::1
Password:
PING ::1(::1) 56 data bytes
ping: sendmsg: Operation not permitted
ping: sendmsg: Operation not permitted
ping: sendmsg: Operation not permitted
ping: sendmsg: Operation not permitted
--- ::1 ping
I'm still frustrated by this freeze. Looking at this thread dump, I can see it
must be the main thread that's hanging - but how can I work out anything more
regarding the problem?
I've done all I can think of to open up the machine it's running on - stopping
all other unnecessary processes
I'd think it's more likely to be timing out while resolving a hostname via
DNS, but starting tomcat via strace should show you exactly where it is
pausing.
On Thu, May 22, 2008 at 5:08 AM, Adam Hardy [EMAIL PROTECTED]
wrote:
I'm still frustrated by this freeze. Looking at this thread dump, I
Found it! But I think my knowledge of linux system calls and kernel functions is
letting me down now. Presumably that is port 48669, which should be open - but
is it a DNS call? If it is a DNS call, then the DNS server on the same machine
should serve it - piping it out to the net, presumably.
Can you compare that output to a capture from the box where you don't see
the pause? I wonder if it's a difference in IPv4/IPv6 configuration between
the boxes...
On Thu, May 22, 2008 at 8:24 AM, Adam Hardy [EMAIL PROTECTED]
wrote:
Found it! But I think my knowledge of linux system calls and
I think this is the equivalent part of the strace log from another machine on
the LAN:
socket(PF_INET6, SOCK_STREAM, IPPROTO_IP) = 10
listen(10, 1) = 0
getsockname(10, {sa_family=AF_INET6, sin6_port=htons(54792), inet_pton(AF_INET6,
::, sin6_addr), sin6_flowinfo=0,
I'm assuming the pause you saw was at the following line?
connect(11, {sa_family=AF_INET6, sin6_port=htons(48669),
inet_pton(AF_INET6, ::1, sin6_addr), sin6_flowinfo=0, sin6_scope_id=0},
28) = -1 ETIMEDOUT (Connection timed out)
That looks like a connect attempt to the IPv6 loopback address (?)
Tell you the truth, I can only assume IPv6 stuff is configured to the debian
etch defaults - although that might have changed between the date I installed
the no-pause-box and the date I installed the problem box.
no-pause-box:
eth0 Link encap:Ethernet HWaddr 00:10:DC:79:FF:8F
I've been trying to solve this problem for a couple of hours now and I can't see
any solution on google or in the archives.
I have tomcat 6.0.13 and java 1.5.0_12 running on debian linux 2.6, my own
installation, not the debian packages.
I stripped the tomcat config down to the bare minimum,
Adam Hardy wrote:
I've been trying to solve this problem for a couple of hours now and I
can't see any solution on google or in the archives.
I have tomcat 6.0.13 and java 1.5.0_12 running on debian linux 2.6, my
own installation, not the debian packages.
I stripped the tomcat config down
just do two thread dumps during the freeze
kill -3 tomcat process id
sleep 5
kill -3 tomcat process id
the thread dump will identify where it is hanging, and it gets output to
std.out, by default to catalina.out
Filip
Adam Hardy wrote:
I've been trying to solve this problem for a couple of
Thanks for the advice. Here are the two thread dumps, with a 5 sec pause
between.
It looks like it's waiting on a network connection. Is that what you would
deduce?
19-May-2008 23:04:32 org.apache.tomcat.util.modeler.BaseModelMBean preRegister
FINE: preRegister [EMAIL PROTECTED]
14 matches
Mail list logo