Comment 8 for bug 1808010
Revision history for this message
| Matt Riedemann (mriedem) wrote Re: Tempest cirros boots fail due to lack of disk space | #8 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
0e1f616
(Get the code!)
(11:25:40 AM) clarkb: frickler: mriedem melwitt to TL;DR the cirros ssh issues, they all root cause down to metadata being slow, but in the case of running out of disk we are attempting to use config drive first (which would ignore the metadata server), but since disk is full that fails then we fall back to metadata then that times out and fails?
(11:26:16 AM) mriedem: metadata api being slow?
(11:26:25 AM) mriedem: as in the guest doesn't get network info fast enough from the meta api?
(11:26:46 AM) clarkb: mriedem: yup, cirros has a 10 second timeout on that network request, and nova was taking 10.something seconds to respond
Which sounds related to bug 1799298 and just general inefficiencies in the metadata-api service serving requests.