-
|
After growing our grid to a large number of nodes, we have discovered a multitude of problems. This one (along with many of our problems) stems from our salt master being extremely busy. With highstates taking excessively long to complete, it might take 2 hours before the sensor shows green again and has all containers running again. Looking for a work around, another guy came up with just doing a docker start on the containers. This brought another thing to mind... Earlier this week, I used a bandaid on some of our zeek containers, Well, we came up with a faster way. We |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
|
Thank you for these observations. Other large grids run without these difficulties, so we suspect that this may be unique to your environment (perhaps due to STIGs, other standards, or Salt customizations). |
Beta Was this translation helpful? Give feedback.
-
|
After fighting through some stubborn so-zeek containers' failures today I may have come to the "why" behind restart not being set on all the containers. A simple docker restart (on the stubborn cases) usually just puts all the zeek processes/workers back into a crashed state. In this case, I had to restart, then execute Perhaps this was the reasoning behind the decision? I'm hoping to find an interim solution (band-aid), but I have too many options at the moment and won't be in my environment until next week... |
Beta Was this translation helpful? Give feedback.
-
|
After fighting through some stubborn so-zeek containers' failures today I may have come to the "why" behind restart not being set on all the containers. A simple docker restart (on the stubborn cases) usually just puts all the zeek processes/workers back into a crashed state. In this case, I had to restart, then execute zeekctl deploy to get them running again. Perhaps this was the reasoning behind the decision? I'm hoping to find an interim solution (band-aid), but I have too many options at the moment and won't be in my environment until next week... |
Beta Was this translation helpful? Give feedback.
After fighting through some stubborn so-zeek containers' failures today I may have come to the "why" behind restart not being set on all the containers. A simple docker restart (on the stubborn cases) usually just puts all the zeek processes/workers back into a crashed state. In this case, I had to restart, then execute zeekctl deploy to get them running again.
Perhaps this was the reasoning behind the decision?
I'm hoping to find an interim solution (band-aid), but I have too many options at the moment and won't be in my environment until next week...