hey,
we’ve got a strange problem on loading the submit script round here. firing up the script on some machines really takes a long time (about 65 seconds), other machines are fine (about 5 sec). testet both with after effects and nuke.
i tried to lock the error down by analyzing the SubmitAEToDeadline.jsx script and figured out that “deadlincommand.exe -pools” takes a long time to get the data (about 21 secs). same with -groups and -getmaximumpriority.
specs:
windows 7
Deadline Version: 7.1.2.1 R (53013d5bd)
FranticX Version: 2.3.0.0 R (c3f0982ed)
any idea?
regards,
jakob
// congaz visual media company GmbH
Since installing Deadline on those machines, have they had a machine restart? Do you have any network/system monitoring tools in place, to confirm it wasn’t network/machine congestion at the MongoDB/Repo machine end when these ‘troublesome’ machines displayed this issue? IS it always, the same few machine(s) which are slow, or does the issue seem to jump around between machines?
yes, the machines have been restartet and yes, the problem occurs always on the same machines.
i first encountered the problem earlier this week, just because i had to sit in front of one of those “slow” machines and did some (still ongoing) testing with the help of our IT. all machines have the same network-config (DHCP DNS etc); but it turns out that deadlinecommand works fast without an internet connection (static ip, leaving the default gateway blank). but again - the “fast” machines does have internet while running the script…
Hi,
Can you try a few things here. Can you flush the dns cache on this machine and try again? Can you also flush the cache on a machine which is fast and try again here as well? (You will need to be admin here)
ipconfig /flushdns
noip.com/support/knowledgeba … dns-cache/
Do you connect to Deadline Repository via UNC path? Can you try mapping this UNC path to a drive letter, then in Deadline Launcher, changing the repo path to this mapped drive letter and then testing again?
i did some more testing (btw thanks for your help!)
flushdns didn’t affect the speed (neither the slow nor the fast).
mapped “\nas\DeadlineRepository7” to “Z:” and changed the path; but it is still slow.
on the slow machines, loading the panels for machine list, the limit groups and the dependencies also takes him a while, but when i submit an ae-project via the monitor, these panels are way more responsive.
If I understand you correctly as per your previous reply, removing the gateway on the slow machine, makes this machine go fast, which to me, points firmly at a routing/networking issue (IT) on this machine. Does this slow machine have more than 1 x network connection / card configured in its “Network Connections” compared to the “fast” machines? The slow deadlinecommand.exe response could well be it trying to connect to your repo/db via 1 network card, fails to resolve (times out 20 seconds later) and then uses another network connection/card and works just fine (1 second).
Deadline Monitor submission will be fast as the network connection has already been resolved. I bet initial startup of Monitor application is slow though…
Perhaps, if you do have multiple network connections, its just a case of configuring the adapter priority:
howtogeek.com/howto/27994/ho … n-windows/
support.microsoft.com/en-us/kb/2526067
thanks for your help. i finally got it.
first thing i tried earlier this week was deactivating windows firewall, antivirus-software etc (…without success). while playing around with the network settings, i realized all of the slow machines running antivirus-software AVG. Because deactivating doesn’t solve the problem, i tried to uninstall AVG.
bam.
seems that AVG AntiVirus FREE 2015 has some kind of control over the internet connection.
Again, thanks for your help.
Cool, glad you got it all sorted! Hmmm…Anti-virus - must remember that one for the next time. 