Harald Milz schrieb:
Hi,
got another issue with 1.3.27. Whenever epg.data is larger than about 6 MBytes, vdr will not start reliably. It will attempt to read epg.data, sit for a while, then restart, and loop.
Is there a reliable way of truncating epg.data if it gets too large?
I set UpdateChannels now to 1, too many useless channels anyway.
My epg.data is about 15MB, and VDR 1.3.27 works perfectly. Are you sure, that is the problem?
Greets
skiller2k1
skiller2k1 skiller2k1@gmx.net wrote:
My epg.data is about 15MB, and VDR 1.3.27 works perfectly. Are you sure, that is the problem?
Technically speaking, I can't be sure because I did not strace vdr. But this is exactly the point of time when vdr dies reproducibly. So, phenomenologically speaking, yes. Even Olaf Titz had a similar problem (See thread "EPG data growing without bounds" from earlier this month).
Harald Milz hm@seneca.muc.de wrote:
skiller2k1 skiller2k1@gmx.net wrote:
My epg.data is about 15MB, and VDR 1.3.27 works perfectly. Are you sure, that is the problem?
Technically speaking, I can't be sure because I did not strace vdr. But
Forgot to add that if I remove epg.data or truncate it to zero before starting vdr the problem goes reproducibly away, on 2 different machines. Hence my question "is there a way to safely truncate epg.data to a certain (safe) length".
I'll try some measurements asap, i.e. which epg.data sizes make vdr crash (?).
On Wed, 2005-07-20 at 08:41 +0200, Harald Milz wrote:
Harald Milz hm@seneca.muc.de wrote:
skiller2k1 skiller2k1@gmx.net wrote:
My epg.data is about 15MB, and VDR 1.3.27 works perfectly. Are you sure, that is the problem?
Technically speaking, I can't be sure because I did not strace vdr. But
Forgot to add that if I remove epg.data or truncate it to zero before starting vdr the problem goes reproducibly away, on 2 different machines. Hence my question "is there a way to safely truncate epg.data to a certain (safe) length".
I'm seeing some EPG problems that could possibly be related; no crashes or kills here, but after running for some time, my VDR is developing an off-by-one (or two) problem with associating programs and their descriptions. The shown description for a program may be the one of the next (or previous, can't remember now) program. This is particularly annoying when later burning such a recording to a DVD using the burn plugin...
I think I've tried all EPG bugfix levels, to no avail. Removing epg.data and restarting VDR appears to fix it. And I don't think I've ever seen the off-by-one problems until VDR writes epg.data to disk some time after a "fresh" restart. Anyone seeing similar effects?
I think I've tried all EPG bugfix levels, to no avail. Removing epg.data and restarting VDR appears to fix it. And I don't think I've ever seen the off-by-one problems until VDR writes epg.data to disk some time after a "fresh" restart. Anyone seeing similar effects?
I'm seeing this occasionally. I think this has something to do with short-term programming changes - perhaps the stations are not always providing EPG data which properly overrides the old versions, or VDR sometimes ignores the new version. The latter could be a VDR bug but the former may be really hard to hunt down...
Olaf
Technically speaking, I can't be sure because I did not strace vdr. But this is exactly the point of time when vdr dies reproducibly. So, phenomenologically speaking, yes. Even Olaf Titz had a similar problem
...and fixed it(*), but I don't think by now that this is the same problem.
You can easily test for that one, however: look in the epg.data file for repeated chunks of lines starting with X. A single event (E line) should not be followed by several identical X lines.
Olaf
(*) http://www.linuxtv.org/pipermail/vdr/2005-July/003466.html
I demand that Olaf Titz may or may not have written...
I think I've tried all EPG bugfix levels, to no avail. Removing epg.data and restarting VDR appears to fix it. And I don't think I've ever seen the off-by-one problems until VDR writes epg.data to disk some time after a "fresh" restart. Anyone seeing similar effects?
I'm seeing this occasionally. I think this has something to do with short-term programming changes - perhaps the stations are not always providing EPG data which properly overrides the old versions, or VDR sometimes ignores the new version. The latter could be a VDR bug but the former may be really hard to hunt down...
I've seen that today on BBC1 wrt a few detonators being set off in London. Basically, what's happening is that the new EPG data (extra news programmes) is being added, but the programmes which are completely replaced aren't removed, at least if they don't have the same start times as the new programmes.