On a stable system that has been installed for 2 years, and no intentional changes have been made, I am having the following issue: (said command below worked as recently as last week).
econe-run-instances --access-key=/tmp/x509up_u2904 --secret-key=/tmp/x509up_u2904 --user-data=timmtest --type=glideinwms.4cpu --url https://fclheadgpvm01.fnal.gov:8444/ ami-00000005
econe-run-instances: select(): Operation now in progress
Both the client and server software come from OpenNebula 4.8. The following statements are true:
- I get this error with all econe-* commands
- It’s known to be a very low-level socket error, signifying that you cannot open the socket
- The error persists against two different servers
- the same econe-run-instances command that is run on the server itself works
- the econe-server is running normally and can still launch virtual machines when accessed via other clients
- I installed a new machine with client software and I get the same error
- reboot of the client machine doesn’t help, neither does restart of the econe-server
- The only difference I can see is that the version of the “curb” ruby gem is different between client and server.
Client
-bash-4.1$ gem list curb
*** LOCAL GEMS ***
curb (0.9.2, 0.9.1, 0.9.0, 0.8.8)
-bash-4.1$
Server
[root@fclheadgpvm01 ~]# gem list curb
*** LOCAL GEMS ***
curb (0.8.8)
Could the version of the curb gem be the issue? If so is there a way to force econe-* utilities to use the older version, and to back-install said older version where I don’t have it?