Opened 14 years ago

Closed 14 years ago

Last modified 13 years ago

#1404 closed defect (worksforme)

Invalid file error at file change with LiveTV

Reported by: bobnvic@… Owned by: Isaac Richards
Priority: minor Milestone: unknown
Component: mythtv Version: head
Severity: medium Keywords:
Cc: cpinkham, adrian.wilkins@… Ticket locked: no

Description

I have three systems, a MBE/FE, SBE/FE and standalone FE, all running svn 9114. The SBE writes to an NFS share located on the MBE machine. I set the standalone FE to Watch LiveTV this morning with it streaming from the tuner in the SBE. It worked very well for a couple hours then at 12:30 I received "RingBuf?(/var/mythtv/tv/1025_20060226123000.mpg): Invalid file (fd 17) when opening '/var/mythtv/tv/1025_20060226123000.mpg'. 12 retries remaining." which then counted down to 0 retries. The file was present on the server, however it was only 337kB. The shows afterward appear to have been recorded without a problem. The frontend went in to prebuffer pauses and 'ignoring livetv eof' statements then continuous statements of 'LiveTV forcing JumpTo? 1' (and two more lines) which caused the log file to grow to 1.7GB in two hours. I have attached the entire backend log and the relevant portion of the frontend log. The backend log was started with -v record and the frontend with -v playback. Please let me know if you need anything else.

Attachments (2)

mythbackend.txt (39.0 KB) - added by bob@… 14 years ago.
mythfrontend.txt (131.1 KB) - added by bobnvic@… 14 years ago.

Download all attachments as: .zip

Change History (6)

Changed 14 years ago by bob@…

Attachment: mythbackend.txt added

Changed 14 years ago by bobnvic@…

Attachment: mythfrontend.txt added

comment:1 Changed 14 years ago by cpinkham

Resolution: invalid
Status: newclosed

I'm assuming that since you're using NFS on your SBE, that you're also using it on your standalone FE. If so, this issue is related to nfs file attribute caching, you need to set actimeo=0 on your mount options on the frontend. Also, this should have been asked in -users first since you would have gotten this same answer there. I'm closing the ticket.

comment:2 Changed 14 years ago by bobnvic@…

Resolution: invalid
Status: closedreopened

Your assumption is incorrect. As stated, the frontend is streaming from the SBE, the NFS share is not mounted on the frontend. The slave backend is writing to an NFS share which is mounted with actimeo=0.

comment:3 Changed 14 years ago by cpinkham

Resolution: worksforme
Status: reopenedclosed

Closing this ticket after discussion on the -dev list where the issue appears related to other non-Myth activity going on on the network/fileserver at the same time which caused the SBE to hang trying to switch from the old recording to the new recording when the program changed.

comment:4 Changed 13 years ago by adrian.wilkins@…

Cc: cpinkham adrian.wilkins@… added

Have to say, I've just experienced this.

Differing cirumstances ; this was FE/BE combo, FE reading files straight off a local disk. At some point this evening it filled the 10GB of remaining disk space on the drive with /var/log with "LiveTV forcing JumpTo? 1" and then sat there with the CPU pegged at 100% for an hour (I just happened to SSH into the box from my hotel room, and wondered why it was so damn slow).

Even if it's down to other things happening, it shouldn't behave like this.

Note: See TracTickets for help on using tickets.