Dear all,
I'm using trafficserver 4.2.3 with cluster 3 node, recently, the cluster is not
stable at peak time, at that time, http request rate to each node is about 4k
req/s , the load of server (CPU, RAM, Disk) is not the problem, could any one
help me about this issue? How can I troubleshoot this?
When it happen, the cluster loss cache mode, that make my backend servers
overload.
The diags.log show up the message below:
[Jul 19 21:32:12.776] Server {0x2b2bc8706700} NOTE: Network congestion to
[10.10.10.7] encountered, reverting to proxy only mode[Jul 19 21:32:13.729]
Server {0x2b2bc8807700} NOTE: Network congestion to [10.10.10.7] encountered,
reverting to proxy only mode[Jul 19 21:32:14.581] Server {0x2b2bc8908700}
NOTongestionE: Network congestion to [10.10.10.7] encountered, reverting to
proxy only mode[Jul 19 21:32:14.881] Server {0x2b2bc8a09700} NOTE: Network
congestion to [10.10.10.7] encountered, reverting to proxy only mode[Jul 19
21:34:13.421] Server {0x2b2bc8706700} NOTE: Network congestion to [10.10.10.7]
cleared, reverting to cache mode[Jul 19 21:34:14.690] Server {0x2b2bc8807700}
NOTE: Network congestion to [10.10.10.7] cleared, reverting to cache mode[Jul
19 21:34:15.424] Server {0x2b2bc8908700} NOTE: Network congestion to
[10.10.10.7] cleared, reverting to cache mode[Jul 19 21:34:15.820] Server
{0x2b2bc8a09700} NOTE: Network congestion to [10.10.10.7] cleared, reverting to
cache mode[Jul 19 21:46:57.825] Server {0x2b2bc3f2e700} NOTE: Network
congestion to [10.10.10.7] encountered, reverting to proxy only mode[Jul 19
21:48:58.532] Server {0x2b2bc3f2e700} NOTE: Network congestion to [10.10.10.7]
cleared, reverting to cache mode
Thanks alot and have a nice day,
An.