updateNetworkACLItem with partialupgrade=false leaves a rule broken when cidrlist or traffictype is left out, and drops an ICMP type given with tcp
维护者通常 1 天内回复
评估
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 新手友好度
- 42/100
- Issue 类型
- 缺陷
- 描述清晰度
- 描述清楚
- 活跃度
- 活跃
- 技术栈
- java
调研方向
Start by reading updateNetworkACLItem, transferDataToNetworkAclRulePojo(), NetworkACLItemDaoImpl.update() and NetworkACLItemCidrsDaoImpl.updateCidrs(); trace validation alongside createNetworkACL and the default partial upgrade path. Check the existing tests for network ACL rule updates and add coverage for missing required fields, invalid traffic type, ICMP parameters with another protocol, and rollback when CIDR updates fail. Done means invalid calls fail before leaving a partial update and valid calls preserve the documented behavior.
由索引模型根据 Issue 内容生成。
描述
problem
updateNetworkACLItem with partialupgrade=false replaces every field of the rule with what the call passes. Two cases go wrong.
1. A field left out is written as null, and the rule is left broken.
When cidrlist or traffictype is left out, transferDataToNetworkAclRulePojo() sets them to null. NetworkACLItemDaoImpl.update() then writes the rule row and only afterwards updates the CIDRs, outside a transaction:
- with
cidrlistleft out,NetworkACLItemCidrsDaoImpl.updateCidrs()fails withCannot invoke "java.util.List.size()" because "cidrs" is null. The call fails, but the row has already been committed with the null traffic type. - from then on
listNetworkACLsfor that list fails withCannot invoke "...TrafficType.toString()" because ... getTrafficType() is null, and applying the list fails the same way inSetNetworkACLCommand, until the rule is updated again with every field.
When protocol is left out, the call fails earlier with a NullPointerException in updateIcmpCodeAndTypeFullUpgrade(). Nothing is written in that case, but the error is an NPE rather than a message. (#14046 removes that NPE; the rule would then be saved without a protocol, which createNetworkACL never allows.)
action left out is already handled: it becomes Allow, as on create.
2. An ICMP type or code given with another protocol is silently dropped.
partialupgrade=false protocol=tcp icmptype=8 is accepted, and the rule is saved as tcp without the ICMP type. createNetworkACL with the same parameters, and the same update with the default partialupgrade=true, both refuse it with "Can specify icmpCode and icmpType for ICMP protocol only".
versions
4.20 and later for the dropped ICMP type; the null fields since partialupgrade was added in 4.12. Seen on 4.22 and main.
The steps to reproduce the bug
- In a VPC, create an ACL list and a rule on it:
createNetworkACL aclid=<list> protocol=tcp startport=80 endport=80 cidrlist=10.1.2.0/24 action=allow traffictype=ingress. updateNetworkACLItem id=<rule> partialupgrade=false protocol=udp startport=53 endport=53. It fails with thecidrs is nullNPE.listNetworkACLs aclid=<list>now fails with thegetTrafficType() is nullNPE.updateNetworkACLItem id=<rule> partialupgrade=false protocol=tcp cidrlist=10.1.2.0/24 action=allow traffictype=ingressrepairs it.- On an ICMP rule,
updateNetworkACLItem id=<rule> partialupgrade=false protocol=tcp icmptype=8 cidrlist=10.1.2.0/24 action=allow traffictype=ingressis accepted, and the rule is tcp with no ICMP type.
What to do about it?
- With
partialupgrade=false, refuse the call whenprotocol,cidrlistortraffictypeis missing, before anything is written. Defaulting them as create does would silently widen a rule to every address, or turn an egress rule into an ingress rule. - Refuse an invalid
traffictypeon update, as create does, instead of ignoring it. - Keep an ICMP type or code given with another protocol, so that the existing validation refuses it.
- Write the rule row and its CIDRs in one transaction, so that a failure on the CIDRs can't leave the rule half updated.
- 主要语言
- Java
- 星标
- 3.1k
- 派生
- 1.4k
- 平均合并
- 7 天 16 小时
- 30 天内合并 PR
- 15
环境准备
- 没有 Dockerfile 或 Docker Compose 文件
- 有 Pull Request 模板
- 阅读贡献指南
从这里开始
- 先读完整个 Issue,再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库,在一个分支上完成修改。
- 提交 Pull Request,并在描述里引用这个 Issue 编号。
apache/cloudstack 的其他 Issue
-
resizeVolume of a running system VM's volume is accepted but the VM never sees the new size可能已有人在做 @bhouse-nexthop 于 3 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 72/100
apache/cloudstack#14645 · 1 条评论 ·
维护者通常 1 天内回复
-
VR reports a load balancer change as applied when haproxy.cfg cannot be replaced可能已有人在做 @bhouse-nexthop 于 4 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 78/100
apache/cloudstack#14633 · 1 条评论 ·
维护者通常 1 天内回复
-
restartNetwork reports a ResourceUnavailableException as an unexpected error (530), unlike restartVPC (534)可能已有人在做 @bhouse-nexthop 于 4 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 78/100
apache/cloudstack#14631 · 1 条评论 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 68/100
apache/cloudstack#14625 · 1 条评论 ·
维护者通常 1 天内回复
-
Remote access VPN users are disconnected on every LB, ACL, port forward or firewall change (xl2tpd restarted on every iptables pass)可能已有人在做 @willbennett-devops 于 12 天前认领。 未关闭
难度 2/5 1-3 小时 新手友好度 72/100
apache/cloudstack#14545 · 1 条评论 ·
维护者通常 1 天内回复
查看 apache/cloudstack 的全部 Issue
相似的 Issue
-
[FEATURE] Expose external refresh, node-level translog and warmer stats可能已有人在做 关联的 PR 仍在进行中或已合并。 未关闭untriaged
难度 2/5 1-3 小时 新手友好度 68/100
opensearch-project/opensearch-prometheus-exporter#680 ·
维护者通常 2 天内回复
-
难度 2/5 1-3 小时 新手友好度 85/100
-
难度 2/5 1-3 小时 新手友好度 66/100
维护者通常 2 天内回复
-
bug
难度 1/5 1 小时以内 新手友好度 78/100
-
Good First Issue
难度 1/5 1-3 小时 新手友好度 88/100
hiero-ledger/hiero-sdk-java#3059 ·
维护者通常 1 天内回复