[GitHub] [trafficcontrol] rawlinp opened a new issue #3705: Creating an invalid/broken Riak cluster in TO can cause requests to get stuck in infinite loop

[GitBox <gi...@apache.org>]

rawlinp opened a new issue #3705: Creating an invalid/broken Riak cluster in TO can cause requests to get stuck in infinite loop
URL: https://github.com/apache/trafficcontrol/issues/3705
 
 
   If RIAK servers are created in TO and set to ONLINE, they will form a TO-internal shared Riak cluster. If no server in this TO-internal Riak cluster can successfully execute a Riak command (e.g. if TO cannot connect to the Riak servers at all), then TO API requests for endpoints that execute Riak commands (e.g. `/api/1.3/cdns/name/:cdn/dnsseckeys.json`) will cause goroutines to get stuck in an infinite loop. These goroutines can each peg a CPU on the TO server and can only be stopped by either restarting TO or possibly fixing the riak servers so that they are able to execute commands.
   
   The riak-go-client could be patched to fix this issue but seems to be currently unmaintained: https://github.com/basho/riak-go-client. There is an open PR that might solve the issue, but has not even been commented on by maintainers: https://github.com/basho/riak-go-client/pull/96.
   
   This issue could be alleviated by checking connectivity to the riak servers before creating the TO-internal shared riak cluster.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services