Hi All We are using OpenSolaris 200906 on V20Z hardware.
We are developing an HPC grid provisioning system for OpenSolaris and are running into problems with system services failing due to race conditions. The system is using installadm to provide automated install of OpenSolaris to systems that run a post install configuration using services. On repeated attempts these post install services are failing because network services are not yet up. We have attempted to set dependencies to make sure that these network services are up first before running our service. During this process of analyzing what is going wrong we are also seeing that OpenSolaris services are failing because there dependencies are not yet available. We believe that problems in the installer system installadm and dhcp are the cause of the initial delay which then snowballs from there. I've attached the system messages log. Our service is the tortuga application mentioned on lines 176 and 178. You should also notice that on line 175 mDNSResponder fails. On line 173 the ethernet interface is just coming up. Before the ethernet interface is up, on line 172 routed is failing. This should not be running before the interface is available in my opinion. On line 170 the dhcpagent is failing to bind. Anyway the system startup will fail and drop in to maintenance mode. Once we have logged into the system via the console. We run svcs -xv and see that the dns service and our tortuga service is in maintenance mode. I disable the dns and reenable the service and it runs fine. The same for our tortuga service and everything works from there. The problem is that this is supposed to be an automated provisioning system and this all has to work automatically without manual intervention. -------------- next part -------------- A non-text attachment was scrubbed... Name: messages Type: application/octet-stream Size: 18108 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/caiman-discuss/attachments/20091027/187f2216/attachment.obj> -------------- next part -------------- -------------- next part -------------- A non-text attachment was scrubbed... Name: 1.jpg Type: image/jpeg Size: 3120 bytes Desc: not available URL: <http://mail.opensolaris.org/pipermail/caiman-discuss/attachments/20091027/187f2216/attachment.jpg> -------------- next part -------------- Bruce Rothermal Email: bruce.rothermal at sun.com Skype: bruce.rothermal Google Talk: bruce.rothermal at gmail.com
