Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
harmeet
@harmeet23466044_twitter
that only MiqSnmp.trap_v2 command queue is not running
Daniel Berger
@djberg96
"Loading production environment (Rails 5.0.7.2)" is not a valid console command, is that what you're asking about?
Keenan Brock
@kbrock
Is cleaning up bundles as simple as rm ~/.gem/ruby/2.5.5/bundler/gems/* ; bundle update
or does bundler get confused when you delete files out from underneath it?
feel like I'm constantly cleaning up ~30G from this directory
Adam Grare
@agrare
@kbrock bundle clean is the "nice" way of doing it, I've also just rm -rf the whole bundler/gems dir and re-run bundle update and everything is fine
Keenan Brock
@kbrock
thnx
Nick LaMuro
@NickLaMuro
@kbrock I think that the UI team is aware of this, but my guess is most of this is from the manageiq-ui-classic clones, since every time you bin/update, a new directory for the gem is created, and the UI assets (aka: npm + yarn + webpack) are put in the gem, not in the main repo
I think Martin H. was trying to look at fixing that, but I my guess is it is a hard problem to solve, and other things take priority
Adam Grare
@agrare
you can also leave ui-classic in your bundler.d overrides so you don't have to re-install the ui assets every bundle update
Gregg Tanzillo
@gtanzillo
@harmeet23466044_twitter The snmp action requires the notifier role be enable on aat lease 1 server in the region. Make sure you have that role enabled and the queue workers should pick uo those messages
Jason Frey
@Fryguy
Yeah @himdel had been trying for a long time... He's currently stuck on typescript not allowing alternate directories
Keenan Brock
@kbrock
sigh - this is why we can't have nice things
Follow the drama 😂
Jason Frey
@Fryguy
Interesting... Seems typescript is gone now 😲
Martin Hradil
@himdel
yup, it should technically be doable now, I just need a big enough chunk of time to start with it again :)
Keenan Brock
@kbrock
or a big enough chunk of energy
Martin Hradil
@himdel
that too :) although with big enough chunk of energy, you can make stuff blow up :D
Keenan Brock
@kbrock
which we're pretty good at
Martin Hradil
@himdel
we're more efficient, we don't need the energy for that :D
Daniel Berger
@djberg96
system calls, blech
tarun agarwal
@tarunkumaragarw_twitter
@gtanzillo we have notifier role is enabled, even in monitoring we are getting alerts, but snmp trap is not generating , because our queue command is not running because evry time we put this command in queue(command - MiqSnmp.trap_v2) but it is not running or dequeuing from queue
@djberg96 no i just asking
Gregg Tanzillo
@gtanzillo
@tarunkumaragarw_twitter Just to be sure I understand, the queue messages are not getting picked up any worker and just remaining in the queue?
tarun agarwal
@tarunkumaragarw_twitter
Whever a real time alert generated it is showing under monitoring menu item, but i setting up.snmp configuration as well it bot genrating snmp trap
Yes for only snmp trap queue messages
They are generating queue messages but still remaining in queue, are not p
@GabrielSVinha
@gtanzillo
But same thing is working in other server , i have tried everything but did not get success on that server
tarun agarwal
@tarunkumaragarw_twitter
M
Gregg Tanzillo
@gtanzillo
@tarunkumaragarw_twitter Only 1 server in the region can have the notifier role active, regardless of whether it’s enabled on multiple servers. But still, one of the server (the one that currently has the role active) should pick it up.
It may be possible that the generic workerd on the server with the notifier role active if busy deliverin gother messages of that same or higher priority
tarun agarwal
@tarunkumaragarw_twitter
@gtanzillo so in one region if we have multiple server , then only in one server will get notifier work ?
Gregg Tanzillo
@gtanzillo
Yes, that’s correct
tarun agarwal
@tarunkumaragarw_twitter
But we are getting alerts on multiple server in the same region , but only snmp trap is not generating for those alerts
Gregg Tanzillo
@gtanzillo
The other alerts may not be tied to the notifier role. It may be helpful to see the current state of the queue messages. You can use this grep on the server with the notifier role active to get the breakdown
grep 'Queue count for state=\["ready' evm.log
tarun agarwal
@tarunkumaragarw_twitter
So we have 4 server in one region , in all 4 server we have enabled notifier role enable and create one alert and configure snmp trap on all server then which server we will get snmp trap generated or alert will be generated
Current state of the queue message is Ready
Gregg Tanzillo
@gtanzillo
Any server can create the message in the queue. But only a generic worker on the server with the role active can pick up the message
tarun agarwal
@tarunkumaragarw_twitter
So how it decides which servers generic worker will pick the queue message as every server has role enabled
Nick Carboni
@carbonin

@tarunkumaragarw_twitter

Only 1 server in the region can have the notifier role active, regardless of whether it’s enabled on multiple servers

Here @gtanzillo is saying that there is a difference between a server that has the role active and one that is only assigned (like a standby, in case a server fails).
So to answer your question, the workers on the server with the role active will pick up the message

I think you can see active vs assigned on the servers diagnostic page
tarun agarwal
@tarunkumaragarw_twitter
@carbonin, @gtanzillo thanks for your help, will see anf let you guys know if need any furhter help
Ko
Greg Procunier
@gprocunier
Can anyone comment on the recent changes Red Hat is doing with cloudforms and how it will affect ManageIQ? Specifically the direction to move development/support for certain features to IBM MCM ?
I had heard that most non Linux integration will be separated, eg Windows / VMWare etc
I am concerned about how that kind of a move would affect smart state analysis on those platforms.
among other things
Oleg Barenboim
@chessbyte
@gprocunier This is not a community question as far as I can tell. If you want to discuss it privately, please ping me offline