[
https://issues.apache.org/jira/browse/OFBIZ-3333?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12852853#action_12852853
]
Jacques Le Roux commented on OFBIZ-3333:
----------------------------------------
Hans,
First some things are worth to be said, in the following
* it leans on a simple POS/MCS configuration (or client/server which is
actually the same)
* the jobs (push and pull) are scheduled and running on the client only
There are actually several knows issues:
# Syncing stay in running status though the related job "crashed". For instance
if an issue occurs (eg, in order of possible apparition: net crash, power
outage on the POS, or server crash for any reason, etc.). This is a small
issue, and I'm not even quite sure it's a design flaw, but it's rather really
designed this way (to keep things under control). Note that I did not design
nor wrote the EntitySunc stuff, so it's only an hypohesis.
** I still wonder if we should really automate the clearing of the running
state (put the status to not startged) when an issue occured, like Pankaj Jain
suggested.
*** This because there is already the "Reset run status" from webtools ->
entity sync status for doing it manyually, so and I don't see the point to
reboot the client for that. And I prefer to keep a hand on that.
*** Else if really needed we could use the underlying
resetEntitySyncStatusToNotStarted service. But we should really take care of
not firing several push or pull jobs. Because by default the jobs are scheduled
to run respectively every 5 minutes and each hour. I suspect the other issues
Pankaj reported ("An instance is already running", "Connection refused") could
be related to this... Also not sure, but it seems it's more a problem on the
client side than for the server, ie more a problem for the push job.
# Some entities miss in the pull syncing. According to Deyan Tsvetanov, this
might also appear during an interruption. This would be really more annoying
and Deyan makes good propositions. But it seems Pankaj never crossed this issue
and I did not neither so far: to be clarified.
# RMI is slow. I think everybody agree on that, but this clearly needs more work
# Pankaj crossed an issue with EECAs. IMO EECAs are delicate beasts and should
be carefully used, but it might be an issue indeed. The idea is to avoid having
duplicated creation on the other end if an EECA is already creating a new
entity on one end. So the EECA should be only on one end.
This is a 1st effort trying to clarify the situation. I hope we can all
cooperate on this and end with a common and shared solution.
> Issues with EntitySync
> ----------------------
>
> Key: OFBIZ-3333
> URL: https://issues.apache.org/jira/browse/OFBIZ-3333
> Project: OFBiz
> Issue Type: Bug
> Components: framework, specialpurpose/pos
> Reporter: Jacques Le Roux
> Assignee: Jacques Le Roux
>
> This is an umbrella task to groups all issues related to the synchronization
> process.
> A subtask will be created for each identified issue.
> I selected framework and POS as affected components, because for the moment
> the synchronization process is mostly used to sync POS terminals. But it
> could be used for other needs.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
