Maybe somebody can help. I have 2 machines, both are p4 2.4ghz with 512MB ram on a 100Mbit network.
I have one as a master backend, it has a pvr-250 in it. The other I use as a frontend, it has no capture card in it. Both are running rh9 with almost identical configs.
My problem lies in communication between the mythfrontend client on the machine without the pvr-250 and the mythbackend on the one that does.
When I go to the watch recordings screen or the delete recordings screen (both show the small preview) any number of odd things happen. Moving around with the arrow keys, once something I recorded is selected the preview will start to play usually but when I move to another recording mythfrontend freezes. Sometimes it's frozen indefinitely and dumps a core, others it shows the preview after almost a minute (rare). During such pauses there is no network activity between the machines and the mythfrontend and mythbackend procs remain either idle or real low on the top list.
The output from mythfront end in these cases varies from:
"Connection timed out," to "could not find codec parameters: myth://backend:6543/file.nuv," to sometimes saying "Couldn't read file "myth://backend:6543/file.nuv"
If I go to the play recordings screen and quickly select a recording it will begin to play after a long while but when I escape from it I also get the slowness and an eventual core dump. In these cases mythfrontend will say "Changing from WatchingPreRecorded to None" or "Changing from None to None" before it dies which I assume is just normal output.
I don't get it. It seems like myth:// requests to my backend die off for the most part. What's weird is if I am in the middle of one of the mentioned pauses, if I switch from the desktop running mythfrontend to another one and then back I can move the arrows again until I move over a different recording, but I sometimes see a frozen frame of the preview of whatever recording I selected that caused the pause.
I'm just wondering if anybody else has had this. Both boxes are running 0.10, the slow down does not seem to be network related because I get the same thing even with machines on a cross-over cable.
One thing to maybe note is the recordings I'm trying to preview are in mpeg2 from the pvr-250, 640x480 at 16000 kb/s (I have alot of disk and figured I'd change it if it ever got low). Is that part of the problem? The files are too big? I get no sign of that from looking at the low resources used on both boxes when trying to see a preview and the non-existent network activity.
Well that's about it. Maybe somebody has a couple suggestions? Or using nfs could fix this in some way? I'm open to everything. Sorry if this may be a common topic, I looked in the list archives and didn't see anything on time outs for months.
Thanks,
Eric
---------------------------------
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
I have one as a master backend, it has a pvr-250 in it. The other I use as a frontend, it has no capture card in it. Both are running rh9 with almost identical configs.
My problem lies in communication between the mythfrontend client on the machine without the pvr-250 and the mythbackend on the one that does.
When I go to the watch recordings screen or the delete recordings screen (both show the small preview) any number of odd things happen. Moving around with the arrow keys, once something I recorded is selected the preview will start to play usually but when I move to another recording mythfrontend freezes. Sometimes it's frozen indefinitely and dumps a core, others it shows the preview after almost a minute (rare). During such pauses there is no network activity between the machines and the mythfrontend and mythbackend procs remain either idle or real low on the top list.
The output from mythfront end in these cases varies from:
"Connection timed out," to "could not find codec parameters: myth://backend:6543/file.nuv," to sometimes saying "Couldn't read file "myth://backend:6543/file.nuv"
If I go to the play recordings screen and quickly select a recording it will begin to play after a long while but when I escape from it I also get the slowness and an eventual core dump. In these cases mythfrontend will say "Changing from WatchingPreRecorded to None" or "Changing from None to None" before it dies which I assume is just normal output.
I don't get it. It seems like myth:// requests to my backend die off for the most part. What's weird is if I am in the middle of one of the mentioned pauses, if I switch from the desktop running mythfrontend to another one and then back I can move the arrows again until I move over a different recording, but I sometimes see a frozen frame of the preview of whatever recording I selected that caused the pause.
I'm just wondering if anybody else has had this. Both boxes are running 0.10, the slow down does not seem to be network related because I get the same thing even with machines on a cross-over cable.
One thing to maybe note is the recordings I'm trying to preview are in mpeg2 from the pvr-250, 640x480 at 16000 kb/s (I have alot of disk and figured I'd change it if it ever got low). Is that part of the problem? The files are too big? I get no sign of that from looking at the low resources used on both boxes when trying to see a preview and the non-existent network activity.
Well that's about it. Maybe somebody has a couple suggestions? Or using nfs could fix this in some way? I'm open to everything. Sorry if this may be a common topic, I looked in the list archives and didn't see anything on time outs for months.
Thanks,
Eric
---------------------------------
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!