[ 
https://issues.apache.org/jira/browse/HDDS-16255?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Meng-Shuan Tsai updated HDDS-16255:
-----------------------------------
    Description: 
ContainerMapper.parseOmDB iterates every OmKeyLocationInfoGroup of every key in 
the DEFAULT/non-FSO keyTable and calls createLocationList() on each, only to 
walk the resulting List once with a for-each and read getBlockID() from each 
entry:
{code:java}
for (OmKeyLocationInfoGroup keyLocationInfoGroup : 
keyInfo.getKeyLocationVersions()) {
  List<OmKeyLocationInfo> keyLocationInfo = keyLocationInfoGroup
      .createLocationList();
  for (OmKeyLocationInfo keyLocation : keyLocationInfo) {
    ...
  }
}
{code}
createLocationList() is documented as not O(1); it flatten-copies each group's 
locationVersionMap into a throwaway List. The call site scans its
selected keyTable without a per-call batching limit; the aggregate allocation
impact relative to other callers requires measurement.

  was:
ContainerMapper.parseOmDB iterates every OmKeyLocationInfoGroup of every key in 
the DEFAULT/non-FSO keyTable and calls createLocationList() on each, only to 
walk the resulting List once with a for-each and read getBlockID() from each 
entry:
{code:java}
for (OmKeyLocationInfoGroup keyLocationInfoGroup : 
keyInfo.getKeyLocationVersions()) {
  List<OmKeyLocationInfo> keyLocationInfo = keyLocationInfoGroup
      .createLocationList();
  for (OmKeyLocationInfo keyLocation : keyLocationInfo) {
    ...
  }
}
{code}
createLocationList() is documented as not O(1); it flatten-copies each
group's locationVersionMap into a throwaway List. The call site scans its
selected keyTable without a per-call batching limit; the aggregate allocation
impact relative to other callers requires measurement.


> Avoid the per-group list copy in ContainerMapper.parseOmDB
> ----------------------------------------------------------
>
>                 Key: HDDS-16255
>                 URL: https://issues.apache.org/jira/browse/HDDS-16255
>             Project: Apache Ozone
>          Issue Type: Sub-task
>            Reporter: Meng-Shuan Tsai
>            Assignee: Meng-Shuan Tsai
>            Priority: Major
>              Labels: pull-request-available
>
> ContainerMapper.parseOmDB iterates every OmKeyLocationInfoGroup of every key 
> in the DEFAULT/non-FSO keyTable and calls createLocationList() on each, only 
> to walk the resulting List once with a for-each and read getBlockID() from 
> each entry:
> {code:java}
> for (OmKeyLocationInfoGroup keyLocationInfoGroup : 
> keyInfo.getKeyLocationVersions()) {
>   List<OmKeyLocationInfo> keyLocationInfo = keyLocationInfoGroup
>       .createLocationList();
>   for (OmKeyLocationInfo keyLocation : keyLocationInfo) {
>     ...
>   }
> }
> {code}
> createLocationList() is documented as not O(1); it flatten-copies each 
> group's locationVersionMap into a throwaway List. The call site scans its
> selected keyTable without a per-call batching limit; the aggregate allocation
> impact relative to other callers requires measurement.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to