[ 
https://issues.apache.org/jira/browse/GEODE-8240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17133774#comment-17133774
 ] 

Bill Burcham commented on GEODE-8240:
-------------------------------------

We see the test bouncing VM 2 to the current version:

{code}
[info 2020/06/11 14:44:37.744 PDT <main> tid=0x1] Bouncing 2 old pid is 19653 
and version is 1.12.0
…
[info 2020/06/11 14:44:44.529 PDT <main> tid=0x1] Bounced 2.  New pid is 19744 
and version is 10240.0.0
{code}

{{10240.0.0}} is secret code for _current version_.

Right after that we see that we are indeed running the latest locator code:

{code}
[vm2] Product-Name: Apache Geode
[vm2] Product-Version: 1.14.0-build.0
{code}

At this point VMs 0 and 1 have not yet been rolled to the current version.

The new locator (VM 2) finds the current coordinator, which appears to be 
running the old version, as expected:

{code}
[vm2] [info 2020/06/11 14:44:45.247 PDT <RMI TCP Connection(1)-192.168.1.27> 
tid=0x12] received 
FindCoordinatorResponse(coordinator=192.168.1.27(vm0:19651)<v1>:41002(version:GEODE
 1.12.0), fromView=true, viewId=-1, 
registrants=[192.168.1.27(vm2:19744:locator)<ec>:41001], 
senderId=192.168.1.27(vm2:19744:locator)<ec>:41001, network partition detection 
enabled=true, locators preferred as coordinators=true, 
view=View[192.168.1.27(vm2:19653:locator)<ec><v0>:41001(version:GEODE 
1.12.0)|-1] members: [192.168.1.27(vm0:19651)<v1>:41002(version:GEODE 
1.12.0){lead}, 192.168.1.27(vm1:19652)<v2>:41003(version:GEODE 1.12.0)]) from 
locator HostAndPort[bburcham-mbpro.hsd1.or.comcast.net:22878]
{code}

But notice the {{view=View[}} section of that log line: it appears that vm2 is 
running the old version, even though the {{registrants=[}} section seems to 
show that vm2 is running the currenty version we expect!

And we can see vm0 preparing the first view that will contain this new vm2:

{code}
[vm0_v1.12.0] [info 2020/06/11 14:44:45.566 PDT <Geode Membership View Creator> 
tid=0x47] preparing new view View[192.168.1.27(vm0:19651)<v1>:41002|9] members: 
[192.168.1.27(vm0:19651)<v1>:41002{lead}, 192.168.1.27(vm1:19652)<v2>:41003, 
192.168.1.27(vm2:19744:locator)<ec><v9>:41001]
{code}

This is view 9 being originated by vm0 and we can see that it appears that vm2 
is running the _same_ version as the originator (vm0) and that is the old 
version.

And when vm2 receives that view:

{code}
[vm2] [info 2020/06/11 14:44:45.573 PDT <unicast receiver,bburcham-mbpro-17341> 
tid=0x23] received new view: 
View[192.168.1.27(vm0:19651)<v1>:41002(version:GEODE 1.12.0)|9] members: 
[192.168.1.27(vm0:19651)<v1>:41002(version:GEODE 1.12.0){lead}, 
192.168.1.27(vm1:19652)<v2>:41003(version:GEODE 1.12.0), 
192.168.1.27(vm2:19744:locator)<ec><v9>:41001(version:GEODE 1.12.0)]
{code}

we see that the member identifier for vm2 in that record has the old version.

The next place to look is in serialization/deserialization of 
{{MemberIdentifierImpl}} and classes it contains.




> Rolling upgrade fails for Locator
> ---------------------------------
>
>                 Key: GEODE-8240
>                 URL: https://issues.apache.org/jira/browse/GEODE-8240
>             Project: Geode
>          Issue Type: Bug
>          Components: client/server, membership
>            Reporter: Ernest Burghardt
>            Assignee: Ernest Burghardt
>            Priority: Major
>
> as shown in [https://github.com/apache/geode/pull/5224]
>  
>  upgrade from version 1.12 doesn't seem to occur on the Locator 
>  
> testRollServersOnPartitionedRegion_dataserializable  failure results:
> Expecting:
>  <"Member Count : 3
>  Name | Id
>  ---- | 
> -------------------------------------------------------------------------------
>  vm2 | 127.0.0.1(vm2:35019:locator)<ec><v17>:41000(version:GEODE 1.12.0) 
> [Coordinator]
>  vm0 | 10.0.0.111(vm0:35025)<v27>:41001
>  vm1 | 10.0.0.111(vm1:35030)<v29>:41002
>  ">
>  not to contain:
>  <"1.12.0">



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to