-
Bug
-
Resolution: Obsolete
-
Major
-
None
-
JDG 7.0.0 GA, JDG 7.1.0 ER5
-
None
-
None
Tested a rolling upgrade from a 3 node JDG 6.6.0 cluster to a 3 node JDG 7 cluster with a replicated cache. Before the migration this was that distribution of data in the source cluster:
209.132.179.19 | SUCCESS | rc=0 >>
/replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 50001
}
209.132.179.15 | SUCCESS | rc=0 >>
/replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 50001
}
209.132.178.120 | SUCCESS | rc=0 >>
/replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 50001
}
The REPL cache has 50k entries plus 1 for the keyset dump.
After the migration, all entries went into a single node (the one where synchronize-data was called). Final distribution:
209.132.179.33 | SUCCESS | rc=0 >>
replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 0
}
209.132.179.34 | SUCCESS | rc=0 >>
replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 0
}
209.132.179.21 | SUCCESS | rc=0 >>
replicated-cache=utility:read-attribute(name=number-of-entries)
{
"outcome" => "success",
"result" => 50000
}
- blocks
-
-
- Closed
-
