TM sends warning logs to log_location_info instead of log_location_warning

[rawlinp]

I'm submitting a ...

• bug report

Traffic Control components affected ...

• Traffic Monitor

Current behavior:

TM currently sends all warnings to the file handle set in log_location_info.

Expected behavior:

TM should send all warnings to the file handle set in log_location_warning.

Minimal reproduction of the problem with instructions:

Set log_location_info and log_location_warning to two different file handles.

On an extremely slow TM server with an relatively large CDN, make this request:

curl -vs 'http://yourtrafficmonitor.example.com/publish/CacheStats?hc=1&wildcard=1&stats=bandwidth,maxKbps,ats.proxy.process.http.current_client_connections,ats.proxy.process.cache.volume_1.wrap_count,ats.proxy.process.cache.volume_2.wrap_count,ats.proxy.process.cache.volume_1.write.failure,ats.proxy.process.cache.volume_1.write.success,ats.proxy.process.ssl.cipher.user_agent.ECDHE-RSA-AES128-GCM-SHA256,ats.proxy.process.ssl.cipher.user_agent.RC4-SHA,ats.proxy.process.ssl.cipher.user_agent.RC4-MD5,ats.proxy.process.ssl.cipher.user_agent.DES-CBC3-SHA,ats.proxy.process.cache.bytes_total,ats.proxy.process.cache.bytes_used,ats.proxy.process.cache.direntries.total,ats.proxy.process.cache.direntries.used,ats.proxy.process.cache.directory_collision,ats.proxy.process.cache.evacuate.active,ats.proxy.process.cache.evacuate.failure,ats.proxy.process.cache.evacuate.success,ats.proxy.process.cache.frags_per_doc.1,ats.proxy.process.cache.frags_per_doc.2,ats.proxy.process.cache.frags_per_doc.3+,ats.proxy.process.cache.gc_bytes_evacuated,ats.proxy.process.cache.gc_frags_evacuated,ats.proxy.process.cache.hdr_marshal_bytes,ats.proxy.process.cache.hdr_marshals,ats.proxy.process.cache.lookup.active,ats.proxy.process.cache.lookup.failure,ats.proxy.process.cache.lookup.success,ats.proxy.process.cache.ram_cache.bytes_used,ats.proxy.process.cache.ram_cache.total_bytes,ats.proxy.process.cache.ram_cache.hits,ats.proxy.process.cache.ram_cache.misses,ats.proxy.process.http.background_fill_bytes_aborted_stat,ats.proxy.process.http.background_fill_bytes_completed_stat,ats.proxy.process.http.background_fill_current_count,ats.proxy.node.cache.bytes_free'

Or just set serve_write_timeout_ms to 1 in the TM config and make the same request on a smaller CDN. This should produce a warning log, which you'll notice is actually logged to log_location_info instead of log_location_warning.