[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] Failing to start: Debian 8 (assertion error)



Both machines have created /var/lock/condor/shared_port_ad:

[Debian]
MyAddress = "<192.168.5.11:9618?noUDP>"

[Ubuntu]
SharedPortCommandSinfuls = "<192.168.5.41:9618>"
MyAddress = "<192.168.5.41:9618?addrs=192.168.5.41-9618&noUDP>"

*Neither* machine has a directory /var/log/condor/daemon_sock/, which the error message refers to.

You can see that the MyAddress line on the Debian machine does not have 'addrs' in it, which is required (for reasons that aren't worth getting into). You can also see that the Debian file is missing the SharedPortCommandSinfuls entirely. This isn't required, but since that attribute made its first appearance in 8.3.5, I'm wondering if you somehow have an old copy of the shared port daemon lying around. (An installation built from source will probably have placed binaries in a different place than a new .deb package, as I understand it.) That could, potentially, also explain why the shared port daemon wasn't able to sends its keepalive to the master.

If you didn't have the shared port daemon turned on, you can always turn it back off with

USE_SHARED_PORT = FALSE

in the configuration file. If you'd like to use shared port, I'd check to make sure that `condor_config_val SHARED_PORT` points to an up-to-date copy of the shared port daemon.

- Todd