Re: heartbeat problem???
Maxence Rousseau <[email protected]>
| Newsgroups | gmane.linux.highavailability.ultramonkey |
|---|---|
| Message-ID | <[email protected]> |
Hi, I think it's because the two node are declared for the other one as unavailable at the same time. The documentation say in two words : "do everything you can to avoid link failure of all the hearbeat links at the same time (use eth0, eth3, serial link, ...)". If it'snt happen on the sametime, it's working with the auto_fallback option (shutdown one node, power it up (wait 1-2 min)), shutdown down the other one ...). Cheers, Max Le lundi 12 juin 2006 à 23:10 +0800, Jiang a écrit : > Hi, > > Thanks a lot for the reply. > > But I am just wondering in case that heartbeat link is failure, if the > system would back to its normal working status after the heartbeat > link is restored? I mean back to the status of one master one slave. > In my tests, it seemed sometimes it won't, as indicated in my last > message, I got two master running. And we can't offord to restart the > whole system to make HA work again... > > Any ideas? > > Thanks! > > - Jiang - > > > > On 6/12/06, Maxence Rousseau <[email protected]> wrote: > > Hello, > > > > You must never remove the both heartbeat cable, it's just mean that the > > other host is down (if you want really remove them, use eth0 & eth2 too > > in ha.cf)... > > > > That's why the documentation suggest to use a serial link. > > (a crossover cable isn't very far of a serial link) > > > > Cheers, > > > > Le lundi 12 juin 2006 à 21:52 +0800, Jiang a écrit : > > > Hi, > > > > > > I am not sure if this is the right place to post this message... > > > anyway... I hv posted it at HA and LVS' mailing list, but seems no > > > body could answer me... > > > > > > In short, I hv two linux machines used to do LVS with HA (all > > > necessary packages were downloaded from ultramonkey). I have enslaved > > > eth0 and eth2 for data connections, eth1 and eth3 are for heartbeat, > > > they are connecting to the other machine > > > through cross cables. I attached ifconfig result at the end of this email. > > > > > > in ha.cf, I have: > > > *********************************************************************************************** > > > ucast eth1 172.16.1.1 > > > ucast eth3 172.16.2.1 > > > *********************************************************************************************** > > > In my test, when disconnected both heartbeat cables, the two nodes > > > became two masters, but the problem happened when I connected > > > heartbeat cables back, none of the masters were willing to shudown, > > > which resulted two master running. And I also noticed that data began > > > to flush the ports(eth0/eth2), which slowed down the network speed > > > dramaticaly. > > > > > > If anybody experienced the same problem or if I have missed something? > > > I have encounted the problem at least twice in my tests, and I have to > > > restart my server each time. > > > > > > Thanks! > > > > > > ifconfig -a: > > > *********************************************************************************************** > > > bond0 Link encap:Ethernet HWaddr 00:17:A4:3B:79:6C > > > inet addr:10.226.200.211 Bcast:10.226.200.255 Mask:255.255.255.0 > > > UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1 > > > RX packets:10919145 errors:5348955 dropped:0 overruns:0 frame:2 > > > TX packets:5445539 errors:2071 dropped:0 overruns:0 carrier:2071 > > > collisions:252799 txqueuelen:0 > > > RX bytes:720982816 (687.5 Mb) TX bytes:359657582 (342.9 Mb) > > > > > > eth0 Link encap:Ethernet HWaddr 00:17:A4:3B:79:6C > > > inet addr:10.226.200.211 Bcast:10.226.200.255 Mask:255.255.255.0 > > > UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 > > > RX packets:5459572 errors:5348954 dropped:0 overruns:0 frame:1 > > > TX packets:2577877 errors:0 dropped:0 overruns:0 carrier:0 > > > collisions:123370 txqueuelen:1000 > > > RX bytes:360491466 (343.7 Mb) TX bytes:170263908 (162.3 Mb) > > > Interrupt:28 > > > > > > eth1 Link encap:Ethernet HWaddr 00:17:A4:3B:79:6B > > > inet addr:172.16.1.1 Bcast:172.16.1.255 Mask:255.255.255.0 > > > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > > > RX packets:1030 errors:0 dropped:0 overruns:0 frame:0 > > > TX packets:1286 errors:0 dropped:0 overruns:0 carrier:0 > > > collisions:0 txqueuelen:1000 > > > RX bytes:194654 (190.0 Kb) TX bytes:215134 (210.0 Kb) > > > Interrupt:29 > > > > > > eth2 Link encap:Ethernet HWaddr 00:17:A4:3B:79:6C > > > inet addr:10.226.200.211 Bcast:10.226.200.255 Mask:255.255.255.0 > > > UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1 > > > RX packets:5459573 errors:1 dropped:0 overruns:0 frame:1 > > > TX packets:2867662 errors:2071 dropped:0 overruns:0 carrier:2071 > > > collisions:129429 txqueuelen:1000 > > > RX bytes:360491350 (343.7 Mb) TX bytes:189393674 (180.6 Mb) > > > Base address:0x6000 Memory:f7fe0000-f8000000 > > > > > > eth3 Link encap:Ethernet HWaddr 00:11:0A:5E:2C:A9 > > > inet addr:172.16.2.1 Bcast:172.16.2.255 Mask:255.255.255.0 > > > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > > > RX packets:1367 errors:0 dropped:0 overruns:0 frame:0 > > > TX packets:1419 errors:0 dropped:0 overruns:0 carrier:0 > > > collisions:0 txqueuelen:1000 > > > RX bytes:253011 (247.0 Kb) TX bytes:254905 (248.9 Kb) > > > Base address:0x6040 Memory:f7f60000-f7f80000 > > > *********************************************************************************************** > > > > > > Best regards, > > > - Jiang - > > > > > > > > > > > -- > > Ultra Monkey - http://www.ultramonkey.org/ > > To UNSUBSCRIBE, email to [email protected], with a body: > > unsubscribe ultramonkey-users [email protected] > > where "[email protected]" is YOUR email address. > > > > > > > -- > Regards, > Jiang Lei > -- Ultra Monkey - http://www.ultramonkey.org/ To UNSUBSCRIBE, email to [email protected], with a body: unsubscribe ultramonkey-users [email protected] where "[email protected]" is YOUR email address.