Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • May 13 12:47

    antgonza on 2022.05

    (compare)

  • May 13 11:45

    antgonza on master

    fixes after 2022.04 fix errors Merge pull request #3199 from a… and 4 more (compare)

  • May 12 22:21
    charles-cowart commented #3200
  • May 12 22:21

    charles-cowart on dev

    2022.05 improve visual minor CHANGELOG.md changes [ski… and 1 more (compare)

  • May 12 22:21
    charles-cowart closed #3200
  • May 12 22:21
    antgonza synchronize #3200
  • May 12 22:19
    coveralls commented #3200
  • May 12 21:43
    antgonza review_requested #3200
  • May 12 21:43
    antgonza opened #3200
  • May 12 20:09

    charles-cowart on dev

    fixes after 2022.04 fix errors Merge pull request #3199 from a… (compare)

  • May 12 20:09
    charles-cowart closed #3199
  • May 12 20:01
    charles-cowart review_requested #3199
  • May 12 18:51
    coveralls commented #3199
  • May 12 18:25
    antgonza synchronize #3199
  • May 12 15:26
    antgonza opened #3199
  • May 02 15:29
    antgonza opened #3198
  • Apr 26 14:48
    antgonza opened #3197
  • Apr 15 17:32
    antgonza closed #1251
  • Apr 15 17:32
    antgonza commented #1251
  • Apr 15 17:28
    antgonza closed #3186
Daniel McDonald
@wasade
vooodooo...
Jose Navas
@josenavas
are we still having the meeting today? Given this problems I don’t think it is worth to have the meeting and is better to try to solve the issue so we can continue the development
Yoshiki Vázquez Baeza
@ElDeveloper
my pr is passing
would two people be kind enough to review->merge
?
disregard that
tests haven't finished
Jose Navas
@josenavas
@ElDeveloper do you remember why we force to use docker? we where having a problem and we had to move to them
it was something related with using too much memory or something like that right?
Joshua Shorenstein
@squirrelo
I believe it was their system overloading so things would timeout since not enough CPU or memory.
Yoshiki Vázquez Baeza
@ElDeveloper
@josenavas, right, moving to docker made it such that builds would no longer fail due to an excess in memory consumption
and it sucks because we are essentially juggling with pieces back and forth
Jose Navas
@josenavas
yeah… grrrr
Yoshiki Vázquez Baeza
@ElDeveloper
:goat: oh well, at least this is going to let us move forward for the time being
what I am kinda worried about is that pip installing qiime will hang for "some" users who have "whatever" specs on their systems, because files cannot be downloaded from the FTP ste
GOD * DAMNIT!!! there's a failure
I think we are running out of memory
Joshua Shorenstein
@squirrelo
what the hell is using so much memory?
Jose Navas
@josenavas
note that those images also do not have a lot of memory...
Yoshiki Vázquez Baeza
@ElDeveloper
ok, the failures were not memory related
were timeouts
Jose Navas
@josenavas
they’re timeouts
Yoshiki Vázquez Baeza
@ElDeveloper
does anyone want to merge as is?
Jose Navas
@josenavas
which is easier to hit on this images because yoou only have 1 CPU with up to 2 if there is enough free resources
Yoshiki Vázquez Baeza
@ElDeveloper
can you repeat that?
oh ok ok
yeah
Jose Navas
@josenavas
haha ok
Yoshiki Vázquez Baeza
@ElDeveloper
if someone restarted the build, note that it already has a failure
and it's a different one ...
:rage:
Daniel McDonald
@wasade
loosing my mind.
no idea how to actually recreate this issue
@ElDeveloper, is it possible to deploy the travis docker image or whatever was being used locally and see if this manifests?
ahhhhhh, headway
wget works from pando login nodes
sorry
compute nodes
not login nodes
the common config (i think) between the docker image, and the login nodes is likely either NAT or round robin DNS
Daniel McDonald
@wasade
once the server enters pasv, the routing on either pando or docker are getting confused
Jose Navas
@josenavas
hey
I just found this and looks related:
travis-ci/travis-ci#3012
Daniel McDonald
@wasade
and not fwd'ing the data to the right client-side port. i think. tcpdumps don't fully support this as it appears comm is 2-way on the pasv port
that is interesting, thanks
issue there is http
might be related if it is a NAT issue
but i don't think http does the insane port juggling that ftp does
Yoshiki Vázquez Baeza
@ElDeveloper
@wasade, I don't think the docker image is public