OpenVZ Forum


Home » General » Support » OpenVZ problem non stop
OpenVZ problem non stop [message #29842] Sat, 26 April 2008 12:55 Go to next message
maser is currently offline  maser
Messages: 3
Registered: April 2008
Location: London
Junior Member
hi there, I have been using openvz for about a week now, everything was running just fine when i had it installed, now suddenly all virtual servers are unaccessable and you cant ping/ssh to them, while the host is okay. I can still be able to ssh from the box to them, but not from outside, I tried to bind the ips to the host then unbind them and stop and start the vds's and they worked well for an hour then again suddenly all goes offline, you cant access from or to them. It's really strange having them off and on I tried to check my sysctl.conf and nothing wrong with it compared to your guide, they only worked when I used the centos script in network to bind all the virtual server ips to the host then unbind them back, it works for an hour then again goes all offline, I would appreciate it so much if you can help me as I got 22 virtual servers running for my customers and it's so bad to get that up and down and unstability, thank you.

Report message to a moderator

Re: OpenVZ problem non stop [message #29873 is a reply to message #29842] Mon, 28 April 2008 09:39 Go to previous messageGo to next message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
Hello,

could you please describe your problem in more detail.

1. What's your network configuration? Do you use venet of veth?
- ip a l (from HN)
- ip ro l (from HN)
- ip rule l (from HN)
- arp -n (fron HN when you lose an access to CTs)
- sysctl -a | grep forward (from HN when it stops working)
2. Anything strange in logs or in dmesg?
3. You can use "tcpdump" utility to explore packets behavior.
When it stops working could you please try to ping one of your CTs from external machine. In that moment try to use "tcpdump" utility on the external machine and on th HN. What does it output?

Thank You!

Report message to a moderator

Re: OpenVZ problem non stop [message #29886 is a reply to message #29842] Mon, 28 April 2008 16:12 Go to previous messageGo to next message
jyrppa is currently offline  jyrppa
Messages: 7
Registered: January 2008
Junior Member
I have similar problem. My VPS obtain IP address via DHCP. I can ping and ssh my VPS some time and then suddenly I can't. TCPDUMP shows incoming messages to but no replies back. There is no resource shortages or any other error logs. Sametime when I can't connect VPS I still can reach main hardware node.

When the problem is on and I go to VPS locally and give dhclient command it sometimes helps..

Report message to a moderator

Re: OpenVZ problem non stop [message #29911 is a reply to message #29873] Tue, 29 April 2008 18:51 Go to previous messageGo to next message
jyrppa is currently offline  jyrppa
Messages: 7
Registered: January 2008
Junior Member
- ip a l (from HN):
2: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
4: sit0: <NOARP> mtu 1480 qdisc noop
link/sit 0.0.0.0 brd 0.0.0.0
6: eth0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet 212.54.16.183/20 brd 212.54.31.255 scope global eth0
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever
8: br0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet 212.54.16.183/20 brd 212.54.31.255 scope global br0
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever
1: venet0: <BROADCAST,POINTOPOINT,NOARP,UP,10000> mtu 1500 qdisc noqueue
link/void
3: veth101.0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever

- ip ro l (from HN):
212.54.16.0/20 dev br0 proto kernel scope link src 212.54.16.183
212.54.16.0/20 dev eth0 proto kernel scope link src 212.54.16.183
default via 212.54.16.1 dev eth0
default via 212.54.16.1 dev br0

- ip rule l (from HN):
0: from all lookup 255
32766: from all lookup main
32767: from all lookup default

- arp -n (fron HN when you lose an access to CTs)
Address HWtype HWaddress Flags Mask Iface
212.54.16.1 ether 00:05:31:F8:B6:39 C br0

- sysctl -a | grep forward (from HN when it stops working):
error: "Operation not permitted" reading key "net.ipv6.route.flush"
error: "Operation not permitted" reading key "net.ipv4.route.flush"
net.ipv6.conf.veth101/0.forwarding = 0
net.ipv6.conf.eth0.forwarding = 0
net.ipv6.conf.br0.forwarding = 0
net.ipv6.conf.default.forwarding = 0
net.ipv6.conf.all.forwarding = 0
net.ipv6.conf.lo.forwarding = 0
net.ipv4.conf.veth101/0.mc_forwarding = 0
net.ipv4.conf.veth101/0.forwarding = 1
net.ipv4.conf.venet0.mc_forwarding = 0
net.ipv4.conf.venet0.forwarding = 1
net.ipv4.conf.br0.mc_forwarding = 0
net.ipv4.conf.br0.forwarding = 1
net.ipv4.conf.eth0.mc_forwarding = 0
net.ipv4.conf.eth0.forwarding = 1
net.ipv4.conf.lo.mc_forwarding = 0
net.ipv4.conf.lo.forwarding = 0
net.ipv4.conf.default.mc_forwarding = 0
net.ipv4.conf.default.forwarding = 1
net.ipv4.conf.all.mc_forwarding = 0
net.ipv4.conf.all.forwarding = 0
net.ipv4.ip_forward = 0

2. no

3.
from VPS:
tcpdump: WARNING: arptype 65535 not supported by libpcap - falling back to cooked socket
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on venet0, link-type LINUX_SLL (Linux cooked), capture size 96 bytes
21:46:18.001065 IP e212-54-19-183.elisa-laajakaista.fi > e212-54-12-63.elisa-laajakaista.fi: ICMP echo reply, id 41459, seq 213, length 64
21:46:19.012210 IP e212-54-19-183.elisa-laajakaista.fi > e212-54-12-63.elisa-laajakaista.fi: ICMP echo reply, id 41459, seq 214, length 64
21:46:19.999945 IP e212-54-19-183.elisa-laajakaista.fi > e212-54-12-63.elisa-laajakaista.fi: ICMP echo reply, id 41459, seq 215, length 64

from HN:
only ICMP echo requests can be seen but no replies.

So the echo replies are lost between VPS and HN.

Report message to a moderator

Re: OpenVZ problem non stop [message #29927 is a reply to message #29911] Wed, 30 April 2008 09:38 Go to previous messageGo to next message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
Hi,

Thank You for the information!

Could you possibly add some info please:

1. "ip a l" (from inside the VE)
2. "ip ro l" (from inside the VE)
3. "ip ru le" (from inside the VE)
4. When you try to ping your VE from the external node please:
"tcpdump -n -i br0" (on NH)
At the same time
"tcpdump -n -i eth0" (inside the VE) (I guess you have got eth0 interface inside VE)

P.S. You should set an IP address only to br0 interface not to the eth0 interface.

Report message to a moderator

Re: OpenVZ problem non stop [message #29951 is a reply to message #29842] Thu, 01 May 2008 06:06 Go to previous messageGo to next message
jyrppa is currently offline  jyrppa
Messages: 7
Registered: January 2008
Junior Member
Here is the additional info:

# ip a l
1: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
3: venet0: <BROADCAST,POINTOPOINT,NOARP,UP,10000> mtu 1500 qdisc noqueue
link/void
inet 127.0.0.1/32 scope host venet0
5: eth0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:0b:cd:fc:19:ff brd ff:ff:ff:ff:ff:ff
inet 212.54.19.183/20 brd 212.54.31.255 scope global eth0

# ip ro l
192.0.2.1 dev venet0 scope link
212.54.16.0/20 dev eth0 proto kernel scope link src 212.54.19.183
default via 212.54.16.1 dev eth0
default via 192.0.2.1 dev venet0

# ip rule l
0: from all lookup 255
32766: from all lookup main
32767: from all lookup default

(from ve)
# tcpdump -n -i eth0
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes
08:57:37.210148 IP 212.54.8.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 317
08:57:37.212672 IP 212.54.8.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 317
08:57:37.300555 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 27130, seq 43, length 64
08:57:38.301405 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 27130, seq 44, length 64
08:57:39.301362 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 27130, seq 45, length 64
08:57:40.219645 IP 212.54.8.1.67 > 255.255.255.255.68: BOOTP/DHCP, Reply, length 316
08:57:40.302579 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 27130, seq 46, length 64

(from ve)
# tcpdump -i venet0
tcpdump: WARNING: arptype 65535 not supported by libpcap - falling back to cooked socket
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on venet0, link-type LINUX_SLL (Linux cooked), capture size 96 bytes
09:01:42.397499 IP e212-54-19-183.elisa-laajakaista.fi > e212-54-12-63.elisa-laajakaista.fi: ICMP echo reply, id 41969, seq 41, length 64

(from HN)
#tcpdump -n -i br0
09:03:00.403286 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 119, length 64
09:03:01.406447 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 120, length 64
09:03:02.402553 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 121, length 64
09:03:03.400700 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 122, length 64
09:03:04.400841 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 123, length 64
09:03:05.406213 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 124, length 64
09:03:06.402854 IP 212.54.12.63 > 212.54.19.183: ICMP echo request, id 41969, seq 125, length 64

Report message to a moderator

Re: OpenVZ problem non stop [message #29952 is a reply to message #29927] Thu, 01 May 2008 06:10 Go to previous messageGo to next message
jyrppa is currently offline  jyrppa
Messages: 7
Registered: January 2008
Junior Member
>P.S. You should set an IP address only to br0 interface not to the eth0 interface.

Yes, now the configuration looks like:

2: lo: <LOOPBACK,UP,10000> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
4: sit0: <NOARP> mtu 1480 qdisc noop
link/sit 0.0.0.0 brd 0.0.0.0
6: eth0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever
8: br0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet 212.54.16.183/20 brd 212.54.31.255 scope global br0
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever
1: venet0: <BROADCAST,POINTOPOINT,NOARP,UP,10000> mtu 1500 qdisc noqueue
link/void
3: veth101.0: <BROADCAST,MULTICAST,UP,10000> mtu 1500 qdisc noqueue
link/ether 00:0d:9d:96:dd:54 brd ff:ff:ff:ff:ff:ff
inet6 fe80::20d:9dff:fe96:dd54/64 scope link
valid_lft forever preferred_lft forever

same problem still.

Report message to a moderator

Re: OpenVZ problem non stop [message #29953 is a reply to message #29952] Thu, 01 May 2008 09:07 Go to previous messageGo to next message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
Hi,

try to remove

default via 192.0.2.1 dev venet0

inside VE.

Report message to a moderator

Re: OpenVZ problem non stop [message #30002 is a reply to message #29953] Mon, 05 May 2008 09:15 Go to previous messageGo to next message
jyrppa is currently offline  jyrppa
Messages: 7
Registered: January 2008
Junior Member
Thanks a lot maratrus! Your solution worked.

Report message to a moderator

Re: OpenVZ problem non stop [message #30018 is a reply to message #29842] Tue, 06 May 2008 10:57 Go to previous messageGo to next message
l4ndy74 is currently offline  l4ndy74
Messages: 27
Registered: August 2006
Junior Member
why br0 and no eth0? Confused
maratrus any ideas? http://forum.openvz.org/index.php?t=tree&goto=29845& #msg_29845 Sad

[Updated on: Tue, 06 May 2008 11:02]

Report message to a moderator

Re: OpenVZ problem non stop [message #30052 is a reply to message #30018] Wed, 07 May 2008 09:29 Go to previous message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
Hello,


could you please describe your problem in more detail.

1. What's your network configuration? Do you use venet of veth?
- ip a l (from HN and from inside the VPS)
- ip ro l (from HN and from inside the VPS)
- ip rule l (from HN nad from inside the VPS)
- arp -n (fron HN when you lose an access to CTs)
- sysctl -a | grep forward (from HN when it stops working)
2. Anything strange in logs or in dmesg?
3. You can use "tcpdump" utility to explore packets behavior.
When it stops working could you please try to ping one of your CTs from external machine. In that moment try to use "tcpdump" utility on the external machine and on th HN. What does it output?
Smile

Quote:


why br0 and no eth0?



Sorry, I don't understand you?
br0 - it's a bridge

Thank You!

Report message to a moderator

Previous Topic: *SOLVED* Cloning containers/VPS
Next Topic: Could not compile a package agains the ovz kernel!?
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sat Jul 27 02:41:55 GMT 2024

Total time taken to generate the page: 0.03170 seconds
Powered by FUDforum Powered by Virtuozzo Containers