[
https://issues.apache.org/activemq/browse/SM-1280?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gert Vanthienen resolved SM-1280.
---------------------------------
Fix Version/s: 3.2.2
3.3
Resolution: Fixed
After discussing it with Guillaume, it seems likely that the issue is caused by
running out of threads in the Executor, causing the sendSync() in the FTP
poller to block because an FTP sender endpoint further down the road also
needs a thread from the same pool. As a result, all connections in the pool
will get allocated to threads stuck on the sendSync(). Fundamentally, the
solution would be to have the FTP poller use asynchronous messaging instead of
using sendSync() (cfr. SM-1282).
For now, a workaround solution is to allow the connection pool to grow as large
as the thread pool when necessary.
> FTP poller stalls because the connection pool is empty
> ------------------------------------------------------
>
> Key: SM-1280
> URL: https://issues.apache.org/activemq/browse/SM-1280
> Project: ServiceMix
> Issue Type: Bug
> Components: servicemix-ftp
> Affects Versions: 3.2.1
> Reporter: Gert Vanthienen
> Fix For: 3.2.2, 3.3
>
> Attachments: smx-ftp-stall-on-borrow.png
>
>
> While processing a lot of files (several thousands), the FTP poller stalls
> with all threads waiting to borrow an object from the pool. I guess the pool
> has become empty somehow. Prior to this, a few SocketTimeoutException were
> thrown (cfr. also SM-915).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
