da_667Mar 16

dalton agent for zeek 8.x+ isn't running, and I have to figure out why.

dockerisms I hate: in a dockerfile, the run command specified becomes an init-like process. If it ever dies, the container dies. If it doesn't run correctly, the container dies every single second. There's no option to get a shell on the box, unless you change the run statement. It makes me want to set fire to a datacenter.

Show threadda_667
docker run -it --fuck-you --do-i-use-the-container-name --or=do_i_need_to_find_the_fucking_containerid
Mar 16 at 4:31pmWeb
Show threadda_667Mar 16

docker logs

haha f u. The container is rebooting repeatedly and I didn't bother to writing any fucking logs as to why.

Show threadBaikalMar 16
@da_667 The -f in docker logs is short for --fuck-you
Show threadJ4ckMar 16
@Baikal @da_667
Show threadkajer :notverified:Mar 16

@da_667 I had a user come to me because their Google Chrome was constantly throwing ERR_NETWORK_CHANGED and user couldn't do ANYTHING.

Naturally it's the network, wifi is broken.

No, you have a docker container on your laptop looping and the docker0 interface is triggering chrome to be dumb.

... docker!!!

Show threadFritz AdalisMar 16
@kajer @da_667
It's fun when users run Docker and you have a network on 172.17.
Show threadkajer :notverified:Mar 16

@FritzAdalis @da_667

All my previous orgs we used 172.16.0.0/12 as router and tunnel links, so this never really came up thank goodness. I had one org using 172.16.x.x as VPN pools, and that did cause issues, but changing the IP pool was easy enough.

I almost favor amazon's method of using 169.254 for all routing links, but have never tried it outside of AWS.

Show threadFritz AdalisMar 16
@kajer @da_667
For stuff that should never route (containers) I like to use 100.64.0.0/10. Might not work out if you're a carrier though.
Show threadkajer :notverified:Mar 16
@FritzAdalis @da_667 When doing iscsi I am FIRMLY in the camp that a L2 protocol should NEVER be routed. so I use 192.0.2.0/24 for IPv4 and a 2001:db8::/64 for ipv6
Show threadtyzbitMar 16
@da_667 you can override the entrypoint to something like tail -f /dev/null or sleep infinity and then run the entrypoint manually yourself with whatever special invocations you want
Show threadthatnaysayerMar 16
@da_667 for me the question is always: did I give the container a proper container_name, which often seems to be "no".
Show threadrjMar 16
@da_667 oh you forgot --entrypoint for good fuckery