Why is output stopping all outputs routing when a 3rd party server goes down?

[lukessi]

Hi,

I am getting a weird issue. If the syslog server fails, it stops all data being indexed by the default TCP out, and then Splunk fills its buckets and falls over. Am I missing something to set it to continue if it can't connect to a output.

cat outputs.conf

[syslog]

defaultGroup = xxxxx_indexers

[syslog:xxxxx_indexers]

server = xxx.xxx.xxx.xxx:9997

type = tcp

timestampformat = %Y-%m-%dT%T.%S

cat transforms.conf

[mehRouting]

REGEX = .

DEST_KEY = _TCP_ROUTING

FORMAT = xxx_cluster_indexers

[Routing_firewalls]

SOURCE_KEY = MetaData:Sourcetype

REGEX = (fgt_traffic|fgt_utm)

DEST_KEY = _SYSLOG_ROUTING

FORMAT = xxxx_indexers

cat props.conf

[host::xxxxxxx1c]

TRANSFORMS-routing = mehRouting, Routing_firewalls

[host::xxxxxc]

TRANSFORMS-routing = mehRouting, Routing_firewalls

[harsmarvania57]

Hi @lukessi,

Can you please provide your full configuration from outputs.conf because I can't see xxx_cluster_indexers in your outputs.conf

[lukessi]

confirmed if I loss the 3rd party syslog, it stops forwarding to our indexers as well.

[harsmarvania57]

Based on answer https://answers.splunk.com/answers/556975/how-to-forward-data-from-an-indexer-to-a-3rd-party.html , it looks like known issue when you'll send data to syslog over TCP. Splunk stops sending data to indexers as well when syslog server is down, better to switch it to UDP or raise case with splunk support.

[lukessi]

Ah yes its picked up from the default index app we have...

[tcpout]

useACK = true

defaultGroup = xxx_cluster_indexers

disabled = false

[tcpout:xxx_cluster_indexers]

server = index1:9997,index2:9997