major routing problem [message #26944] |
Wed, 06 February 2008 03:37 ![Go to next message Go to next message](/theme/ovz3/images/down.png) |
sara3
Messages: 38 Registered: February 2008
|
Member |
|
|
Hello
i have a carpon copy problem of this thread
http://forum.openvz.org/index.php?t=msg&goto=7489&
my problem is as follows
suddenly after a server reboot i cannot get any apache inside my cpanel ves to start
i checked error_log of apache for each ve they had
[notice] caught SIGTERM, shutting down
i suspected apache and rebuilt it on all ves
and after 48 hours of continious sleepless tries
i found out the cause of the problem by strace the httpd process inside one ve
14206 fcntl64(16, F_SETFD, FD_CLOEXEC) = 0
14206 setsockopt(16, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
14206 setsockopt(16, SOL_SOCKET, SO_KEEPALIVE, [1], 4) = 0
14206 setsockopt(16, SOL_TCP, TCP_NODELAY, [1], 4) = 0
14206 bind(16, {sa_family=AF_INET, sin_port=htons(443), sin_addr=inet_addr("0.0.0.0")}, 16) = -1 EADDRINUSE (Address already in use)
14206 time(NULL) = 1202264293
14206 write(2, "[Wed Feb 6 04:18:13 2008] [crit] (98)Address already in use: ma"..., 99) = 99
14206 write(2, "\n", 1) = 1
14206 close(16) = 0
14206 exit_group(1) = ?
i tried to stop httpd service inside hardware node and all ve's apache began to work as normal as before
when i start apache all ve's apache begin to fail again
i want you to kindly help me fixing the issue because i really really want and need httpd running on main hardware node
also httpd was running on hardware node for almost a year with no problems
waiting your replies
|
|
|
|
|
Re: major routing problem [message #27041 is a reply to message #26990] |
Thu, 07 February 2008 11:34 ![Go to previous message Go to previous message](/theme/ovz3/images/up.png) |
den
Messages: 494 Registered: December 2005
|
Senior Member |
|
|
Unfortunately, I have no idea whats happen. It works and stops to work. Seems quite strange.
What can be done:
- upgrade to never kernel. 028stab023 is quite odd
In order to understand the problem, you could do the following:
- start apache in VE0
- dump netstat -ln
- start VE _without_ httpd
- dump netstat -ln inside
- if port 443 is busy, try to locate the listener with lsof (inside VE). If this does not help, stop apache in VE0 and re-run netstat in both environments.
- if port 443 is in VE, install netcat package into VE
- start nc -l -p 443. if netcat is able to start, then the port 443 can be listened in VE on 0.0.0.0
- stop netcat
- try to start httpd manually. If this does not help, obtain full strace of httpd starting attempt and analyze/send it to me. It is the victim
|
|
|