Bug #27635
closed
Salt fails to pull pillar information from foreman
Description
We are noticing that when running salt against many hosts, foreman-node is returning HTTP service errors for a subset of hosts.
Command that we run:
[root@saltmaster ~]# /usr/bin/salt --async 'se*' state.highstate queue=True test=True
Example of the error from /var/log/salt/master:
2019-08-14 22:00:35,432 [salt.loaded.int.module.cmdmod:1113][ERROR ][6442] output: Couldn't retrieve ENC data: Error retrieving node setest1.com: Net::HTTPServiceUnavailable
2019-08-14 22:00:35,433 [salt.loaded.int.module.cmdmod:1113][ERROR ][6504] output: Couldn't retrieve ENC data: Error retrieving node setest2.com: Net::HTTPServiceUnavailable
2019-08-14 22:00:35,433 [salt.loaded.int.module.cmdmod:1113][ERROR ][6399] output: Couldn't retrieve ENC data: Error retrieving node setest3.com: Net::HTTPServiceUnavailable
2019-08-14 22:00:35,433 [salt.loaded.int.module.cmdmod:1113][ERROR ][6520] output: Couldn't retrieve ENC data: Error retrieving node setest4.com: Net::HTTPServiceUnavailable
Updated by Bernhard Suttner over 4 years ago
- Project changed from Foreman to Salt
Which version of the plugin are you using?
Updated by Joe Collette over 4 years ago
- Status changed from New to Resolved
We were able to resolve this by tuning passenger on the foreman server to accept more connections. EPEL passenger/passenger-mod were also replaced by packages from the foreman repo.
- cat /etc/httpd/conf.d/05-foreman-ssl.conf |grep Pass
PassengerAppRoot /usr/share/foreman
PassengerRuby /usr/bin/tfm-ruby
PassengerMinInstances 6
PassengerStartTimeout 600
PassengerMaxPreloaderIdleTime 0
PassengerMaxRequests 10000
PassengerStatThrottleRate 180
PassengerMaxRequestQueueSize 500
PassengerMaxPoolSize 50
PassengerHighPerformance on
Updated by Bernhard Suttner over 4 years ago
I see, I guess, this issue is then fixed and was not related to foreman_salt.