Page 1 of 5 | Next »
[17:07:47] gheegh: so, i have a question guys.. i'm super confused.. I have a system that uses sidekiq. When I launch it.. i get 100% failure rate with a NoMethodError. The error comes from an old version of a gem. When i Rails console and run the same job with the same arguments, it doesn't have a problem (it uses the right version of the gem). Why would sidekiq not be using the same gems as rails?
[17:08:01] gheegh: so, i have a question guys.. i'm super confused.. I have a system that uses sidekiq. When I launch it.. i get 100% failure rate with a NoMethodError. The error comes from an old version of a gem. When i Rails console and run the same job with the same arguments, it doesn't have a problem (it uses the right version of the gem). Why would sidekiq not be using the same gems as rails?
[17:08:40] gheegh: i have literally gone into my machine and deleted all the previous versions of gems in various folders
[17:10:17] gheegh: what's also weird, is that these report to Airbrake.. and AirBrake says the file location is /bundler/gems/. But, there's no /bundler folder on this machine.
[17:12:57] gheegh: sorry.. #ruby is my go to normally.. after i posted, i realized it was probably a rails question
[17:15:04] gheegh: it's a really strange thing.. i'm thoroughly confused.. i've spent several hours trying to figure out why my fixes weren't working, and realized it's still using an old version
[17:31:35] gheegh: dang Tmux. I had closed some sessions wiht workers in them.. and it didn't kill the process.. that was why.
[11:05:22] gheegh: file itself already processed, but I guess that's deprecated now? Any thoughts on how to eliminate the IO and preprocessing?
[11:05:22] gheegh: hey all. Question. I have a gem i wrote that i'm tyring to improve the speed of. Basically, every time it runs, it loads its config file from disk (0.04s) and deserializes data (0.04) taking almost a 10th of a second. Not that big of a problem, except that this is called HUNDREDS of times per each process. This ends up adding almost 10 seconds to a job process. I was looking at adding the data to the bottom of the config.rb
[12:07:48] gheegh: right now, i'm just throwing them into hashes that are in the source code itself.. which is a bit clumsy.. but it might work.. if that owrks, then i'll generate a rake/bin type task that takes the config data and genreates a class that has this data in it.. so we can update the config in a nice yaml format and then produce the cached/computed file..
[12:20:15] gheegh: its got a module_function method that is the main "callable" called .parse and this pulls data from the config when it needs it
[20:48:04] gheegh: Hey all.. question. I have a problem where my sidekiq processes just seem to grow and grow and grow.. I've spent a week or so profiling, and i think the problem boils down to that we do retrievals, and when we get several big ones at once, the ruby process consumes more memory. .and never releases it. So, i'm wondering, has anyone ever FORKED in the middle of a sidekiq process (to wrap the retriever part), so i can then
[14:19:56] gheegh: Hey all.. I'm wondering if anyone knows of any guide on best practices for jobs that "retrieve data"? I have a system that's doing hundreds of thousands of jobs.. occasionally.. my Sidekiq job gets locked up.. and then re-queues itself, then it just seems to cascade.. what i need is a clean Timeout that will, after a fixed period of time, re-queue to a problem queue. Over the course of a day, it will lock up all my worker
[14:33:42] gheegh: yeah, but Sidekiq writes all over their code base not to use Timeout because it's not thread safe.
[14:34:09] gheegh: to timeout the underlying process, retrieve, etc. which i've done.. but there are still some case where it locks the thread..
[00:18:00] gheegh: hey all, anyone have a suggestion as to how to "memoize" a dynamic singleton class variable?
[00:19:55] gheegh: I've got a folder that has a bunch of files in it that i'm converting into class methods on a class. I just realized that one of hte problems is that its reloading them from disk every time it's called, versus just 1 time. I need to "memoize" them.. but i'm not sure exactly how to do that. Here's a gist.. imagine the files as something that i'm turning into Arrays..
[00:45:51] gheegh: so, this works.. its just that it seems to reload the method everytime it's called
[18:57:00] gheegh: Anyone work with active_admin? I've got a problem with a custom form. It's not rendering well, and fields are not showing. Rails 5.2 latest everything. https://gist.github.com/wflanagan/5f73eab4dfdd93aadf74d98fa4c62ba5
[19:02:06] gheegh: maybe I'll ask here, as it's maybe busier.. i have an ActiveAdmin issue. My custom form is totally not working.
[19:02:31] gheegh: It's not rendering well, and fields are not showing. Rails 5.2 latest everything. https://gist.github.com/wflanagan/5f73eab4dfdd93aadf74d98fa4c62ba5 gist shows a link to a pic, and the rendered HTML (and source of form)
[13:06:16] gheegh: hey all.. i need some thoughts. I'm trying to dynamically define a class as a child of OpenStruct, if it's not defined. If defined, I want to just use it. Basically, i'm trying to wrap OpenSturct for responses to an API call. I'm getting a nameerror when trying to use my defined class. thoughts on how to do this?
[13:07:46] gheegh: sorry, to follow the guide, here's the code snippet defined at dpaste.de https://dpaste.de/OEJz#L
[21:03:29] gheegh: Hey all, i'm having a problem with VCR in some of my tests.. i am doing the query in the "let", but it seems to retrieve each time no matter what.. any thoughts on how to fix this in RSpec?
[21:17:23] gheegh: hey all.. anyone using VCR? I'm having some problem figuring out what to do to get my "let" in RSPEC not re-retrieve the page..
[21:18:41] gheegh: https://gist.github.com/wflanagan/59f0434d11079ace97ed4679d503d31d this is an example.. not sure what i'm doing wrong.. the URL is identical.
[21:47:44] gheegh: yeah, but i would have thought vcr would have matched it.. and returned the value
[19:49:31] gheegh: hey all.. question. I'm using Moneta to bulid a simple gem that looks up/loads a file that's about 800MB uncompressed. I am storing it on Github, so I need to break it into smaller file sizes. I had loaded it into several small SQL files, but that's not supported on Heroku. The File store breaks each key into its own file, adding about 400,000 files to the github repo (not a great idea for performance). I need something smaller,
[16:26:43] gheegh: Hey all, anyone have a good, reliable config for using StatsD with ActiveJob? It seems my stats are not getting sent reliably.
[15:38:39] gheegh: essentailly, I'll be checking the database into the repo.. so the problem with DBM is it's machine dependent.
[15:57:49] gheegh: I would love to be thinking in those terms.. too much going on right now.. i'm elbow deep in todos.. :-)