From: Daniel Browne <dbrowne@(email surpressed)> Subject: Render nodes failing to pick up jobs after new license push Date: Mon, 28 Jun 2010 22:06:19 -0400 |
Msg# 1944 View Complete Thread (2 articles) | All Threads Last Next |
Hi Greg, After we upgraded our license a week ago, we occasionally found that some render machines would not pick up new jobs. In particular we encountered this issue with machines that had been recently onlined, either manually during the day or by our idle time script in the evenings. Restarting the rushd daemon on the individual machines seems to resolve the issue, however we puzzled as to why this occurred in the first place. Your instructions for license installation indicate that we only needed to restart the Rush daemon on the Rush server, but not on any of the client machines. Did I misread your instructions or could this be a separate issue? -Dan |
From: Greg Ercolano <erco@(email surpressed)> Subject: Re: Render nodes failing to pick up jobs after new license push Date: Tue, 29 Jun 2010 02:51:50 -0400 |
Msg# 1945 View Complete Thread (2 articles) | All Threads Last Next |
Daniel Browne wrote: > After we upgraded our license a week ago, we occasionally found that some render > machines would not pick up new jobs. In particular we encountered this > issue with machines that had been recently onlined, either manually during > the day or by our idle time script in the evenings. Restarting the rushd > daemon on the individual machines seems to resolve the issue, however we > puzzled as to why this occurred in the first place. Your instructions for > license installation indicate that we only needed to restart the Rush daemon > on the Rush server, but not on any of the client machines. Did I misread > your instructions or could this be a separate issue? Hi Daniel, It sounds like you read the instructions correctly; you would only restart the license server daemon, not the others. And even that isn't entirely necessary; you can either restart the license server daemon, or just wait after installing the license.dat file until the rushd.log shows the daemon has reloaded the license before pushing out the new license to the rest of the machines. (This usually takes up to 30 seconds before the 'Reloading' message appears in the rushd.log) What you're describing here sounds similar to the problem you reported back in December 2009: http://seriss.com/cgi-bin/rush/newsgroup-threaded.cgi?-viewthread+1907+1908+1909+1910+1911+1912+1913+1914+1915 The symptoms being: after a license or hosts file is changed and pushed out, the machines continue to respond OK to 'rush -ping' but stop showing up in rushtop(1), and won't respond in 'rush -lac' (All Cpus) or 'rush -laj' (All Jobs) reports. If the problem you're encountering is on Macs running Leopard or Snow Leopard, make sure the version of rushd (ie. 'rushd -version') reports "102.42a9d" on those machines. If not, please upgrade those Macs to the latest 102.42a9d release: http://www.seriss.com/rush/releases/Upgrade-102.42a9c.html (This page is password protected; notify me by email directly if you don't know your Rush download password.) Note that even though that page is the 102.42a9c release (emphasis on 'c'), the Mac release had a special fix that bumped the version to 102.42a9d (emphasis on 'd'). If you've currently got 'd' on those machines, then the problem might be something else; maybe you can give me more details then, such as if there are any errors in the rushd.log, and if the machines respond at all to 'rush -ping', or if the daemons (rushd) are even running. -- Greg Ercolano, erco@(email surpressed) Seriss Corporation Rush Render Queue, http://seriss.com/rush/ Tel: (Tel# suppressed) Fax: (Tel# suppressed) Cel: (Tel# suppressed) |