-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Could not fetch specs from https://rubygems.org/ #4189
Comments
@eromoe if you change your source as you said, or also changed the sources in the Gemfile, this error you see is strange. I can not reproduce it and I guess is due to some other hiccups, might be some cache, who knows. As you said in your last comment, not sure why is removed, it works for you. Can you confirm this again? My output doing the same thing as you, but with a local gem repo, is:
as you can see the data is fetched not from rubygems, but from the local gem repo. Closing this for now, feel free to reopen if you can provide more information. |
I am using version 2.1.1 on SLES and still have this issue that even when I use "--local" and "--no-verify" it tries to connect to "https://rubygems.org".
|
@Neo23x0 the right command to install from an offline package would be
after properly installing locally the package generated by `bin/plugin pack`` your command is actually trying to install the gem plugin directly as a local gem, so it try to fetch dependencies from the internet and not locally as expected. I see this as an actual improvement for the plugin manager, can you please open a new issue with it? Thanks a lot! |
Actually trying to install the gem directly was the second try. (and yes unpack runs without errors)
Unpack
logstash-output-syslog-2.1.0.gem is in the root of the .tar.gz |
Interesting, this should obviously not happen. Will investigate.
thanks a lot for your report. |
System: SuSE Enterprise Linux (SLES) 12 (x64) java -version java version "1.7.0" I have sudo rights on bin/plugins but the system administrator executed the commands on my request with the same results. |
@Neo23x0 interesting bug, I just tried with a centos6 and this is what I get (machine completely offline) and using the rpm logstash package.
|
I used a package system, which is Ubuntu as described here: I installed logstash-output-syslog there (online) then packed and unpacked on the target machine (offline), which is SLES 12. On the target machine, logstash-output-syslog should be in the local cache folder after "unpack" and would therefore not produce the error "Plugin not found, aborting" as in your case, right? |
this helps a lot, thanks a lot for your feedback. Danke sehr! I will reproduce. |
De nada ;) |
@Neo23x0 if everything went right, you have to see this:
in your target machine cache directory. if this happen, logstash is expected to pick it up from there and install it without issues. |
this is not relevant to my understanding, if no internet there is no internet. |
@Neo23x0 I'm lost with this issue, just tried all your workflow with same config and OS, no connection to outside world for sure, and
everything installed as expected from local repo. there have to be something else happening here. can you share:
|
I got a directory listing of the folder /opt/logstash/vendor/cache/
|
Oh wait - there is one more thing to mention: |
Is there a fail over to the online download if the local gem cannot be found? |
@Neo23x0 There's one more thing you brought up that may be playing into this (emphasis on may):
Logstash and Elasticsearch do not support the IBM JDK/JRE. I know some users are able to use it, but with mixed results. This adds a level of complexity because it will be hard for us to accurately reproduce your results. |
I agree with you @untergeek, @Neo23x0 this was never tested with IBM jdk, are you able to test with oracles or openjdk? |
So, @purbon you tested on a SLES Linux with Oracle JDK yesterday? |
Do I need the JDK for Logstash? |
@Neo23x0 yes, with oracle jdk and openjdk. Will test today with IBM jdk to see if this is actually the issue, generally speaking is a good guess as we have seen very strange behaviour with this jdk. JRE is enough to run logstash. |
Ha - check this out.
|
Interesting,
On Fri, Jan 22, 2016 at 11:17 AM Florian Roth [email protected]
|
I must say, I've never heard about a host named bundler.rubygems.org :-(
On Fri, Jan 22, 2016 at 11:27 AM Pere Urbon-Bayes [email protected]
|
Damnit ! Connects to |
I'll try to figure out what the actual problem is. |
Works also on production system. BTW other issue that made debugging harder: The plugin installation with "sudo" caused the gem and gemspecs files to be written as "root:root" with no read rights for everyone. This caused logstash that runs as "logstash" user in our environment to ignore the module on initialization printing the "missing plugin" message to stdout. |
Because china block the official source, we use
https://ruby.taobao.org/
to instead.But logstash do not use the source I set:
The text was updated successfully, but these errors were encountered: