Conan 2.0 remote login sometimes fail

Sonatype Nexus Repository
,
1

Versions

  • Self-managed Sonatype Nexus Repository PRO 3.77.1-01 (trial)
  • Java openjdk 17.0.14 2025-01-21
  • PostgreSQL 17.4 (NOTE: We originally used the built-in H2 database. Upgraded to try to fix this issue with no luck)
  • conan 2.0

Issue

  • The command “conan remote login …” works for a while, then stops working for a while, then starts working again.
  • This problem is pretty severe because it’s baked into our CI workflow. Each code checkin will do roughly 3 separate conan logins, then later logout (I tested without logging out, and no difference).
  • When the login stops working, it just hangs for a while, and hits the 30 second timeout, thus failing the whole CI pipeline.

Other Thoughts

  • Conan 2.0 is supposed to be well supported by pro. We’re on trial right now, so I can’t create an official support ticket, but I’m hoping someone here knows whats going on.
  • Is there a limited number of conan logins that Nexus allows? If so, is that configurable?
2

There is no limit by Nexus, I would suggest looking in the Nexus request.log around the time of the problematic login and verify that the request takes that long from the perspective of Nexus (e.g. is something else interfering with the request).

Another possibility is that if you’re using an external authentication provider (e.g. LDAP, Crowd, SAML) perhaps they’re the source of the delay.