Spawn doesn't support flow groups

[NerdyNick]

So there is made reference in the User Guide that the Spawn command
supports a 3rd optional param of a Flow Group. However if you attempt
to execute it with that option you get back an error. Looking at the
Java code it also appears that spawn doesn't even support this option.
Nor does it appear any code has been started to get it in there.

So is this a feature that is coming down the road and maybe it should
be commented out of the User Guide tell ready, or was it an idea that
never got implemented nor will it.?

--
Nick Verbeck - NerdyNick

[Henry Robinson]

Hi Nick - 

Flow ids are assigned by the config command, not the spawn command. The format is:

config logical_node [optional flowid] source sink

Can you point me to where in the documentation you see this? I want to make sure I correct all the instances.

cheers,

Henry

--
Henry Robinson
Software Engineer
Cloudera
415-994-6679

[NerdyNick]

Not sure the file but its under "Logical Configurations" -> "Flow
Isolation" right after "Figure 1: Flume Flows: Single Flow"

----------------------------------------------------
NerdyNick.com
Coloco.ubuntu-rocks.org

[KC Braunschweig]

On Aug 2, 12:25 pm, Henry Robinson <he...@cloudera.com> wrote:
> Flow ids are assigned by the config command, not the spawn command. The
> format is:
>
> config logical_node [optional flowid] source sink

Maybe I'm dense, but can you give an exact syntax example? I'm having
trouble getting this to work. Tried both the web multi-config
interface and the shell. Example:

user_event.collector.hostname1 user_event.flow : collectorSource |
collectorSink("hdfs://hostname1/log_archive/dev/%Y/%m/%d/%
{host}/","user_event.%H.", 60000);

Returns:

com.cloudera.flume.conf.FlumeSpecException: Parser error: unexpected
'collectorSource' at position 49 line 1:
'user_event.collector.hostname1 : user_event.flow collectorSource |
collectorSink("hdfs://hostname1/log_archive/dev/%Y/%m/%d/%
{host}/","user_event.%H.", 60000);'

Thanks,

KC

[KC Braunschweig]

Answered my own question. Here is an example in flume shell syntax:

exec config 'collector1' 'foo.flow' 'collectorSource'
'collectorSink("hdfs://collector1/log_archive/dev/%Y/%m/%d/%
{host}/","user_event.%H.", 60000)'

On Aug 12, 4:18 pm, KC Braunschweig <kcbraunschw...@gmail.com> wrote:
> On Aug 2, 12:25 pm, Henry Robinson <he...@cloudera.com> wrote:
>

> >Flowids are assigned by the config command, not the spawn command. The

[Patrick Hunt]

KC, thanks for the followup. Would you mind creating a JIRA (if you
haven't already). Thanks!

Patrick