[
https://issues.apache.org/activemq/browse/SMX4KNL-88?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=46077#action_46077
]
Jamie Goodyear commented on SMX4KNL-88:
---------------------------------------
After much testing adding in "utils sleep 30" to the beginning of all command
chained servicemix startup commands seems to negate the occurrence of hangs on
a failed sub-command. Users of SMX4 command chaining must be made aware that
this practice of putting the kernel to sleep is the work around for possibly
failing commands.
Hence a chain where we may have a command fail (simulated by 'ecco' command):
[EMAIL PROTECTED]:/x1/fuse/esb/bin$ ./servicemix echo hi1 \; ecco hi2\; echo
hi3 \; exit
we would want to use the chain:
[EMAIL PROTECTED]:/x1/fuse/esb/bin$ ./servicemix utils sleep 30000 \; echo hi1
\; ecco hi2\; echo hi3 \; exit
to avoid having the 'ecco hi2' hang servicemix.
Please note that this particular hang situation tends to occur on slower or
heavy load systems.
> Failed kernel commands in command chain will hang processing.
> -------------------------------------------------------------
>
> Key: SMX4KNL-88
> URL: https://issues.apache.org/activemq/browse/SMX4KNL-88
> Project: ServiceMix Kernel
> Issue Type: Bug
> Environment: Unix
> Reporter: Jamie Goodyear
> Assignee: Jamie Goodyear
> Attachments: smx4knl-88a.txt
>
>
> Failed kernel commands in command chain will hang processing.
> Ex 1:
> [EMAIL PROTECTED]:/x1/fuse/esb/bin$ ./servicemix echo hi1 \; echo hi2\; echo
> hi3 \; exit
> The command chain to individually echo "hi1", "hi2" and hi3" will print to
> the terminal screen:
> hi1
> hi2
> hi3
> Then smx will exit, returning the terminal prompt.
> Ex 2:
> [EMAIL PROTECTED]:/x1/fuse/esb/bin$ ./servicemix echo hi1 \; ecco hi2\; echo
> hi3 \; exit
> Here the command chain was intended to individually echo "hi" but in this
> case the second echo command has been mis-spelled. Instead of an error
> message being displayed to the screen and continuing on to echo "hi3" however
> smx stops processing commands (the user then has to ctrl+c the process to
> regain the terminal session).
> Is this the behavior we want to maintain?
> Could we set some property perhaps to allow the kernel to ignore failed
> commands in a chain?
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
