桉树实例将无法启动，立即因停止而失败->待处理->因错误而停止

Question

我正在运行Eucalyptus 4.0-环境良好，已经运行了好几年没有问题。 我经历了关闭过程（停止所有实例，停止eucalyptus-cloud，停止eucalyptus-cc，停止每个节点）并最近关闭环境以进行迁移。

当我恢复环境时，所有服务都重新联机，但没有实例会启动-新的，旧的等。我注意到有关IP分配的一些问题（在此过程中网络没有更改），因此我将所有这些都释放回云中然后重新分配它们。

然后由于观察到的其他错误，我遇到了一些在线信息，最终修改了两个参数：

euca-modify-property -p cloud.network.global_max_network_tag=2048
euca-modify-property -p cloud.network.global_min_network_tag=1024

完成此操作后，我再次重新启动了云，便能够成功启动新实例。 在没有现有实例的情况下，我升级了-> 4.0.1-> 4.0.2。 一切似乎都没有问题地升级了（我的控制台仍然报告4.0.0，但是euca版本报告的是eucalyptus 4.0.2和euca2ools 3.1.1 / Omega）。

但是，我大约需要14个小时的时间，并且无法启动旧的[EBS支持]实例。 它从停止->待处理->停止->在几秒钟内停止-甚至只能从日志中得知。 我相信“ metadata_extant_network”表中还有一些多余的数据（也许某些东西没有正确关闭？），但是我无法确定是什么，也由于FK约束我不能手动删除记录，并且我不想冒险破坏数据库。 这是我尝试启动实例时的日志-必须有一种“适当”的方式来执行此操作...：

cloud-exhaust.log

Tue Dec 9 10:04:29 2014  WARN [org.jboss.netty.channel.DefaultChannelPipeline:Eucalyptus.eucalyptus:Ephemeral
[bitronix.tm.twopc.Preparer:Eucalyptus.eucalyptus:EphemeralConfiguration:arn:euca:eucalyptus:::com.eucalyptus.network.DispatchingNetworkingService/.class java.util.concurrent.ThreadPoolExecutor$Worker#346] executing transaction with 0 enlisted resource
Tue Dec 9 10:04:30 2014  WARN [org.hibernate.engine.jdbc.spi.SqlExceptionHelper:Eucalyptus.eucalyptus:EphemeralConfiguration:arn:euca:eucalyptus:::com.eucalyptus.network.DispatchingNetworkingService/.class java.util.concurrent.ThreadPoolExecutor$Worker#346] SQL Error: 0, SQLState: 23503
Tue Dec 9 10:04:30 2014 ERROR [org.hibernate.engine.jdbc.spi.SqlExceptionHelper:Eucalyptus.eucalyptus:EphemeralConfiguration:arn:euca:eucalyptus:::com.eucalyptus.network.DispatchingNetworkingService/.class java.util.concurrent.ThreadPoolExecutor$Worker#346] ERROR: update or delete on table "metadata_extant_network" violates foreign key constraint "fk6a62681ed068841d" on table "metadata_network_group"
  Detail: Key (id)=(c75a9938419237320141929ac6a02eea) is still referenced from table "metadata_network_group".

postgresql-Tue.log

ERROR:  update or delete on table "metadata_extant_network" violates foreign key constraint "fk6a62681ed068841d" on table "metadata_network_group"
DETAIL:  Key (id)=(c75a9938419237320141929ac6a02eea) is still referenced from table "metadata_network_group".
STATEMENT:  delete from metadata_extant_network where id=$1 and version=$2
ERROR:  update or delete on table "metadata_extant_network" violates foreign key constraint "fk6a62681ed068841d" on table "metadata_network_group"
DETAIL:  Key (id)=(c75a9938419237320141929ac6a02eea) is still referenced from table "metadata_network_group".
STATEMENT:  delete from metadata_extant_network where id=$1 and version=$2

cloud-output.log

2014-12-09 10:04:30 ERROR | org.hibernate.exception.ConstraintViolationException: could not execute statement
2014-12-09 10:04:41  INFO | :1418144681687:Address:ADDRESS_STATE:TOP:Address 192.168.0.216 arn:aws:euare:000000000001:user/nobody available 0.0.0.0  AddressTransition system:unallocated->impending(true)
2014-12-09 10:04:41 ERROR | com.eucalyptus.cloud.util.MetadataException: org.hibernate.LazyInitializationException: could not initialize proxy - no Session
2014-12-09 10:04:41  WARN | Aborting resource token: ResourceToken:i-812D40D4:resources=TypedContext:{com.eucalyptus.util.TypedKey(NetworkResources)=[com.eucalyptus.compute.common.network.PrivateNetworkIndexResource(5), com.eucalyptus.compute.common.network.PublicIPResource()]}

cloud-debug.log

Tue Dec 9 10:04:30 2014 ERROR [NetworkGroups:Eucalyptus.eucalyptus:EphemeralConfiguration:arn:euca:eucalyptus:::com.eucalyptus.network.DispatchingNetworkingService/.class java.util.concurrent.ThreadPoolExecutor$Worker#346] org.hibernate.exception.ConstraintViolationException: could not execute statement
Tue Dec 9 10:04:41 2014  INFO [AdmissionControl:Compute.10] Found authorized clusters: [cc-192.168.0.150]
Tue Dec 9 10:04:41 2014  INFO [AdmissionControl:Compute.10] Availability: cc-192.168.0.150 -> 5
Tue Dec 9 10:04:41 2014 ERROR [ClusterAllocator:Eucalyptus.cluster:ClusterConfiguration:arn:euca:eucalyptus:cluster01:cluster:cc-192.168.0.150/.class java.util.concurrent.ThreadPoolExecutor$Worker#458] com.eucalyptus.cloud.util.MetadataException: org.hibernate.LazyInitializationException: could not initialize proxy - no Session
Tue Dec 9 10:04:41 2014  WARN [Allocations:Eucalyptus.cluster:ClusterConfiguration:arn:euca:eucalyptus:cluster01:cluster:cc-192.168.0.150/.class java.util.concurrent.ThreadPoolExecutor$Worker#458] Aborting resource token: ResourceToken:i-812D40D4:resources=TypedContext:{com.eucalyptus.util.TypedKey(NetworkResources)=[com.eucalyptus.compute.common.network.PrivateNetworkIndexResource(5), com.eucalyptus.compute.common.network.PublicIPResource()]}

cloud-error.log

Tue Dec 9 10:04:30 2014 ERROR [NetworkGroups:Eucalyptus.eucalyptus:EphemeralConfiguration:arn:euca:eucalyptus:::com.eucalyptus.network.DispatchingNetworkingService/.class java.util.concurrent.ThreadPoolExecutor$Worker#346] org.hibernate.exception.ConstraintViolationException: could not execute statement
Tue Dec 9 10:04:41 2014 ERROR [ClusterAllocator:Eucalyptus.cluster:ClusterConfiguration:arn:euca:eucalyptus:cluster01:cluster:cc-192.168.0.150/.class java.util.concurrent.ThreadPoolExecutor$Worker#458] [com.eucalyptus.cloud.run.ClusterAllocator.cleanupOnFailure(ClusterAllocator.java):274] com.eucalyptus.cloud.util.MetadataException: org.hibernate.LazyInitializationException: could not initialize proxy - no Session

因此，然后我直接登录PostgreSQL数据库，删除了FK约束，并手动删除了日志中标识的行：

ALTER TABLE metadata_extant_network DROP CONSTRAINT "fk45157a25f1ac537e";
ALTER TABLE metadata_network_group DROP CONSTRAINT "fk6a62681ed068841d";
DELETE FROM metadata_extant_network WHERE id='c75a9938419237320141929ac6a02eea';

尝试重新启动实例后，删除操作成功完成，我收到一个新错误：

euca-start-instances: error (InternalFailure): Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free.

Tue Dec 9 11:04:23 2014 ERROR [org.mule.exception.DefaultMessagingExceptionStrategy:Compute.15] 
********************************************************************************
Message               : Component that caused exception is: DefaultJavaComponent{Compute.component}. Message payload is of type: StartInstancesType
Code                  : MULE_ERROR--2
--------------------------------------------------------------------------------
Exception stack is:
1. Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free. (com.eucalyptus.cloud.util.NotEnoughResourcesException)
  com.eucalyptus.network.NetworkGroup:325 (null)
2. Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free. (com.eucalyptus.cloud.util.NotEnoughResourcesException)
  com.eucalyptus.cloud.run.AdmissionControl$RunAdmissionControl:148 (null)
3. Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free. (java.lang.RuntimeException)
  com.eucalyptus.util.Exceptions:255 (null)
4. Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free. (com.eucalyptus.util.EucalyptusCloudException)
  com.eucalyptus.compute.service.ComputeService:69 (null)
5. Component that caused exception is: DefaultJavaComponent{Compute.component}. Message payload is of type: StartInstancesType (org.mule.component.ComponentException)
  org.mule.component.DefaultComponentLifecycleAdapter:352 (http://www.mulesoft.org/docs/site/current3/apidocs/org/mule/component/ComponentException.html)
--------------------------------------------------------------------------------
Root Exception stack trace:
com.eucalyptus.cloud.util.NotEnoughResourcesException: Failed to allocate network tag for network: arn:aws:euca:eucalyptus:821881850233:security-group/ownCloud/: no network tags are free.
    at com.eucalyptus.network.NetworkGroup.extantNetwork(NetworkGroup.java:325)
    at com.eucalyptus.network.GenericNetworkingService$_prepareSecurityGroup_closure3_closure12.doCall(GenericNetworkingService.groovy:198)
    at sun.reflect.GeneratedMethodAccessor770.invoke(Unknown Source)
    + 3 more (set debug level logging or '-Dmule.verbose.exceptions=true' for everything)
********************************************************************************

Answer 1

您似乎已为vlan标签配置了与您的安全组设置不兼容的值。 除非需要保留vlan标签用于其他用途，否则不应限制全局范围。

https://www.eucalyptus.com/docs/eucalyptus/4.0.2/#install-guide/configuring_security_groups.html