[ 
https://issues.apache.org/jira/browse/SLING-10528?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Timothee Maret updated SLING-10528:
-----------------------------------
    Description: 
Journal distribution serialise content packages in binary less mode only. When 
doing "deep" distribution, the package can still become so large that the 
importing side can't ingest it. When this happens, Apache Oak does a session 
save that never seem to return and can easily take 8 hours to terminate.

I suggest to detect those large packages, based on the package size and simply 
reject packages above a  configurable limit size. The limit should take in 
consideration the mean import throughput for a single Oak session on cloud 
segment tar and should keep the save operation below, say 15 minutes.

  was:
Currently the package import is logged at INFO level before processing the 
package
{code}
log.info("Importing distribution package {} at offset={}", pkgMsg, offset);
{code}

This makes it difficult to figure out if there was an error with the import 
because one has to seek further logs to see if the package import failed or 
(worst) no log if the package was imported successfully. We should ensure that 
a single log statement is logged upon import, either the import succeeded or 
failed.


> Reject large packages
> ---------------------
>
>                 Key: SLING-10528
>                 URL: https://issues.apache.org/jira/browse/SLING-10528
>             Project: Sling
>          Issue Type: Bug
>          Components: Content Distribution
>            Reporter: Timothee Maret
>            Assignee: Timothee Maret
>            Priority: Major
>             Fix For: Content Distribution Journal Core 0.1.18
>
>
> Journal distribution serialise content packages in binary less mode only. 
> When doing "deep" distribution, the package can still become so large that 
> the importing side can't ingest it. When this happens, Apache Oak does a 
> session save that never seem to return and can easily take 8 hours to 
> terminate.
> I suggest to detect those large packages, based on the package size and 
> simply reject packages above a  configurable limit size. The limit should 
> take in consideration the mean import throughput for a single Oak session on 
> cloud segment tar and should keep the save operation below, say 15 minutes.



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

Reply via email to