Hi,
Appia has a mechanism that chooses the interface to bind in the case
that there are several interfaces in the machine. This was built to
avoid the InetAddress.getLocalHost() that caused problems. This
mechanism was not builded thinking about vlans, so maybe there is an
issue there.
I'm posting this to the appia mailing list so that it can be tested
by the Appia team, but for the moment can you test the following:
Go to org/continuent/appia/protocols/utils/HostUtils.java and change
the variable debug to System.out (I know, we have to change this!)
When you run the Sequoia controller, maybe it will print some output
that will help us to understand why does it choose the wrong interface.
You can also activate log4j hedera appia implementation to debug by
adding this to the log4j.properties file:
log4j.logger.org.continuent.hedera.appia=DEBUG, Console
and post the output to the Appia mailing list.
Thanks for your feedback,
nuno
On Mar 26, 2007, at 4:56 , Sylvain Coutant wrote:
Nuno Carvalho a écrit :
This can happen because of networking problems. TCP would help you
with its flow control. Can you give us details about why the
channels with TCP did not work?
My config on both servers is using VLANs :
eth0 Link encap:Ethernet HWaddr 00:16:3E:00:00:17
inet6 addr: fe80::216:3eff:fe00:17/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:223219144 errors:0 dropped:0 overruns:0 frame:0
TX packets:214197345 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:36519915622 (34.0 GiB) TX bytes:35349115989
(32.9 GiB)
eth0.10 Link encap:Ethernet HWaddr 00:16:3E:00:00:17
inet addr:10.0.0.58 Bcast:10.0.0.127 Mask:255.255.255.128
inet6 addr: fe80::216:3eff:fe00:17/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1496 Metric:1
RX packets:2058381 errors:0 dropped:0 overruns:0 frame:0
TX packets:2337870 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:1679627790 (1.5 GiB) TX bytes:2675119647 (2.4 GiB)
eth0.24 Link encap:Ethernet HWaddr 00:16:3E:00:00:17
inet addr:192.168.0.27 Bcast:192.168.0.31 Mask:255.255.255.248
inet6 addr: fe80::216:3eff:fe00:17/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1496 Metric:1
RX packets:211779752 errors:0 dropped:0 overruns:0 frame:0
TX packets:211859450 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:29579174652 (27.5 GiB) TX bytes:32673990531
(30.4 GiB)
I tried to force the conf to use vlan 24 :
<channel name="TCP SEQ Channel" template="tcp_sequencer"
initialized="yes">
<memorymanagement size="40000000" up_threshold="15000000"
down_threshold="7000000" />
<chsession name="hederalayer">
<parameter name="gossip_address">224.1.1.5</parameter>
<parameter name="local_address">192.168.0.27:12000</
parameter>
</chsession>
</channel>
And both servers try to run on vlan 10 :
Controller Sequoia02 :
17:44:27,579 INFO controller.core.Controller Sequoia controller
(2.10.6)
17:44:27,696 INFO controller.core.Controller Loading configuration
file: /test/persist/sequoia/config/controller/controller.xml
17:44:27,949 INFO controller.core.Controller JMX is enabled
17:44:27,984 INFO controller.core.Controller Starting JMX server
on host: sequoia02.testplatform.com
17:44:28,415 INFO controller.backup.BackupManager Registering
backuper MySQL to handle format MySQL raw dump
17:44:28,440 INFO backend.DatabaseBackend.sqldev3 Adding
connection manager for virtual user "testme"
17:44:28,448 INFO backend.DatabaseBackend.sqldev4 Adding
connection manager for virtual user "testme"
17:44:28,808 INFO controller.RequestManager.sequoia Request
manager will parse requests with the following granularity: TABLE
17:44:28,820 INFO controller.virtualdatabase.sequoia Using Hedera
properties file: /hedera_appia.properties
17:44:29,106 INFO appia.xml.AppiaXML Loading XML configuration
from file: /test/persist/sequoia/config/appia.xml
17:44:29,562 INFO continuent.hedera.appia AppiaThread: Starting
Appia.
17:44:29,602 DEBUG continuent.hedera.gms Appia reported new view:
version: 1
group: [Group:sequoia]
id: [ViewID:0;[AppiaHedera@/10.0.0.58:12000]]
previous: []
view: [[AppiaHedera@/10.0.0.58:12000],]
addresses: [/10.0.0.58:12000,]
17:44:31,608 INFO controller.virtualdatabase.sequoia Group sequoia
connected to Member(address=/10.0.0.58:12000, uid=10.0.0.58:12000)
17:44:31,609 INFO controller.virtualdatabase.sequoia First
controller in group sequoia
17:44:31,668 INFO sequoia.controller.recoverylog Checking recovery
log consistency
17:44:31,677 INFO controller.virtualdatabase.sequoia Enabling
backend sqldev3 from checkpoint disable all backends-
sequoia02.testplatform.com:25322-20070326174247685+0200
17:44:31,985 INFO backend.DatabaseBackend.sqldev3 Detected backend
as: MySQL
17:44:32,105 INFO controller.RequestManager.sequoia Setting new
virtual database schema.
17:44:32,246 INFO controller.recoverylog.RecoverThread Starting
recovery
17:44:32,250 DEBUG continuent.hedera.adapters Sending message 0 to
1 members.
17:44:32,316 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.58:12000, uid=10.0.0.58:12000) for message 0
17:44:32,320 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.58:12000, uid=10.0.0.58:12000) to message 0
17:44:32,320 INFO controller.RequestManager.sequoia All activity
is suspended for sequoia
17:44:32,324 INFO controller.recoverylog.RecoverThread Recovery
completed
17:44:32,366 DEBUG continuent.hedera.adapters Sending message 1 to
1 members.
17:44:32,371 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.58:12000, uid=10.0.0.58:12000) for message 1
17:44:32,374 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.58:12000, uid=10.0.0.58:12000) to message 1
17:44:32,374 INFO controller.recoverylog.RecoverThread Database
backend sqldev3 is now enabled
17:44:32,376 DEBUG continuent.hedera.adapters Sending message 2 to
0 members.
17:44:32,378 WARN controller.virtualdatabase.sequoia No response
while getting static metadata from remote controller
17:44:32,395 INFO controller.virtualdatabase.sequoia Enabling
backend sqldev4 from checkpoint disable all backends-
sequoia02.testplatform.com:25322-20070326174247685+0200
17:44:32,705 INFO backend.DatabaseBackend.sqldev4 Detected backend
as: MySQL
17:44:32,728 INFO controller.RequestManager.sequoia Virtual
database schema merged with new schema.
17:44:32,745 INFO controller.recoverylog.RecoverThread Starting
recovery
17:44:32,748 DEBUG continuent.hedera.adapters Sending message 3 to
1 members.
17:44:32,780 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.58:12000, uid=10.0.0.58:12000) for message 3
17:44:32,783 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.58:12000, uid=10.0.0.58:12000) to message 3
17:44:32,783 INFO controller.RequestManager.sequoia All activity
is suspended for sequoia
17:44:32,788 INFO controller.recoverylog.RecoverThread Recovery
completed
17:44:32,801 DEBUG continuent.hedera.adapters Sending message 4 to
1 members.
17:44:32,805 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.58:12000, uid=10.0.0.58:12000) for message 4
17:44:32,809 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.58:12000, uid=10.0.0.58:12000) to message 4
17:44:32,809 INFO controller.recoverylog.RecoverThread Database
backend sqldev4 is now enabled
17:44:32,813 INFO controller.core.Controller Adding
VirtualDatabase sequoia
17:44:32,821 INFO controller.core.Controller Waiting for
connections on 192.168.0.27:25322
17:44:32,852 INFO controller.core.Controller Controller started on
2007.03.26 13 at 05:44:32 PM CEST
17:44:32,852 INFO controller.core.Controller Controller
sequoia02.testplatform.com:25322 ready, listening to requests ...
Controller Sequoia01 :
17:43:53,815 INFO controller.core.Controller Sequoia controller
(2.10.6)
17:43:53,948 INFO controller.core.Controller Loading configuration
file: /test/persist/sequoia/config/controller/controller.xml
17:43:54,181 INFO controller.core.Controller JMX is enabled
17:43:54,220 INFO controller.core.Controller Starting JMX server
on host: sequoia01.testplatform.com
17:43:54,798 INFO controller.backup.BackupManager Registering
backuper MySQL to handle format MySQL raw dump
17:43:54,860 INFO backend.DatabaseBackend.sqldev1 Adding
connection manager for virtual user "testme"
17:43:54,873 INFO backend.DatabaseBackend.sqldev2 Adding
connection manager for virtual user "testme"
17:43:55,295 INFO controller.RequestManager.sequoia Request
manager will parse requests with the following granularity: TABLE
17:43:55,311 INFO controller.virtualdatabase.sequoia Using Hedera
properties file: /hedera_appia.properties
17:43:55,541 INFO appia.xml.AppiaXML Loading XML configuration
from file: /test/persist/sequoia/config/appia.xml
17:43:56,390 INFO continuent.hedera.appia AppiaThread: Starting
Appia.
17:43:56,469 DEBUG continuent.hedera.gms Appia reported new view:
version: 1
group: [Group:sequoia]
id: [ViewID:0;[AppiaHedera@/10.0.0.57:12000]]
previous: []
view: [[AppiaHedera@/10.0.0.57:12000],]
addresses: [/10.0.0.57:12000,]
17:43:58,477 INFO controller.virtualdatabase.sequoia Group sequoia
connected to Member(address=/10.0.0.57:12000, uid=10.0.0.57:12000)
17:43:58,478 INFO controller.virtualdatabase.sequoia First
controller in group sequoia
17:43:58,535 INFO sequoia.controller.recoverylog Checking recovery
log consistency
17:43:58,551 INFO controller.virtualdatabase.sequoia Enabling
backend sqldev1 from checkpoint disable all backends-
sequoia01.testplatform.com:25322-20070326174258962+0200
17:43:58,969 INFO backend.DatabaseBackend.sqldev1 Detected backend
as: MySQL
17:43:59,136 INFO controller.RequestManager.sequoia Setting new
virtual database schema.
17:43:59,283 INFO controller.recoverylog.RecoverThread Starting
recovery
17:43:59,312 DEBUG continuent.hedera.adapters Sending message 0 to
1 members.
17:43:59,725 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.57:12000, uid=10.0.0.57:12000) for message 0
17:43:59,731 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.57:12000, uid=10.0.0.57:12000) to message 0
17:43:59,732 INFO controller.RequestManager.sequoia All activity
is suspended for sequoia
17:43:59,736 INFO controller.recoverylog.RecoverThread Recovery
completed
17:43:59,773 DEBUG continuent.hedera.adapters Sending message 1 to
1 members.
17:43:59,780 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.57:12000, uid=10.0.0.57:12000) for message 1
17:43:59,785 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.57:12000, uid=10.0.0.57:12000) to message 1
17:43:59,785 INFO controller.recoverylog.RecoverThread Database
backend sqldev1 is now enabled
17:43:59,787 DEBUG continuent.hedera.adapters Sending message 2 to
0 members.
17:43:59,789 WARN controller.virtualdatabase.sequoia No response
while getting static metadata from remote controller
17:43:59,817 INFO controller.virtualdatabase.sequoia Enabling
backend sqldev2 from checkpoint disable all backends-
sequoia01.testplatform.com:25322-20070326174258962+0200
17:44:00,233 INFO backend.DatabaseBackend.sqldev2 Detected backend
as: MySQL
17:44:00,261 INFO controller.RequestManager.sequoia Virtual
database schema merged with new schema.
17:44:00,281 INFO controller.recoverylog.RecoverThread Starting
recovery
17:44:00,326 DEBUG continuent.hedera.adapters Sending message 3 to
1 members.
17:44:00,342 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.57:12000, uid=10.0.0.57:12000) for message 3
17:44:00,363 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.57:12000, uid=10.0.0.57:12000) to message 3
17:44:00,364 INFO controller.RequestManager.sequoia All activity
is suspended for sequoia
17:44:00,368 INFO controller.recoverylog.RecoverThread Recovery
completed
17:44:00,387 DEBUG continuent.hedera.adapters Sending message 4 to
1 members.
17:44:00,394 DEBUG continuent.hedera.adapters Replying to Member
(address=/10.0.0.57:12000, uid=10.0.0.57:12000) for message 4
17:44:00,402 DEBUG continuent.hedera.adapters Received reply from
Member(address=/10.0.0.57:12000, uid=10.0.0.57:12000) to message 4
17:44:00,403 INFO controller.recoverylog.RecoverThread Database
backend sqldev2 is now enabled
17:44:00,407 INFO controller.core.Controller Adding
VirtualDatabase sequoia
17:44:00,416 INFO controller.core.Controller Waiting for
connections on 192.168.0.26:25322
17:44:00,437 INFO controller.core.Controller Controller started on
2007.03.26 13 at 05:44:00 PM CEST
17:44:00,437 INFO controller.core.Controller Controller
sequoia01.testplatform.com:25322 ready, listening to requests ...
I don't want to use vlan 10 because it's a special closed network
running with dhcp addresses that are not suitable for that purpose.
In fact, even if I don't force the IP address in the TCP config,
controllers never see each other.
Also note the 1496 MTU on interfaces that could perhaps be the
cause of some trouble ...
BR,
Sylvain.
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
--
Nuno Carvalho
University of Lisbon, Portugal
http://dialnp.di.fc.ul.pt
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
