search

Strider-CD / strider-node

Run Node.js tests in strider
11 stars 13 forks source link

Crashes strider #20

Open knownasilya opened 9 years ago

knownasilya commented 9 years ago

Not sure how to recreate this, but it seems like during npm install process in prepare phase, if it takes a while, it crashes strider. No errors that I can see.

Have strict caching enabled and node stable.

fernandoneto commented 9 years ago

having same issue and can't found any errors.

knownasilya commented 9 years ago

@fernandoneto one issue i've found is that big processes take more memory, and if your system is low on swap memory, the OOM Killer takes over and kills the process.

fernandoneto commented 9 years ago

im running in a instance with 1 gb of memory. when the npm install process start cpu go to 100% and memory 800/900 :( .

knownasilya commented 9 years ago

@fernandoneto I've solved this by doing builds locally and doing npm install --production instead. This saves on significant overhead for npm install. This might not be feasible for your situation..

fernandoneto commented 9 years ago

@knownasilya sorry but i don't understand your aproach.

knownasilya commented 9 years ago

Meaning that I forgo installing the development dependencies since I build my client assets locally. But this would only work if you don't run tests and only use Strider as continuous deployment to a like an alpha server.

cusspvz commented 8 years ago

@fernandoneto, npm install --production installs only resources we need to initialize a server environment bypassing development dependencies, although we need those to test our environment which isn't a good solution for a CI.

@knownasilya unfortunately we may discard the use of strider since, as @fernandoneto told me personally, doesn't clean up testing containers. :/

knownasilya commented 8 years ago

By testing containers are you referring to the fact that build data doesn't get wiped out?

cusspvz commented 8 years ago

@fernandoneto is in charge of it, and on last crashes we experienced he reported that the system had lots of containers, and probably strider wasn't removing all of them.

EDIT: I meant stopped containers.

knownasilya commented 8 years ago

Like Docker containers? Would love to know more..

cusspvz commented 8 years ago

ping @fernandoneto

fernandoneto commented 8 years ago

@knownasilya yes docker containers, our tests runs inside docker containers. When we facing the last crash i have to reboot the server to can acess over ssh and there are lots of stoped containers. The most of then in the status have exit a couple of minutes ago, so maybe all the container's runnig caused a lot of memory usage and server crash. I can not specify the number of container's stoped but is something around 15,20.