Which version of MythTV are people running?

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Whoa...big about-face here:

I've just been told on the mailing list that using the --nologserver command line option does in fact do console logging ONLY, and the only way to get file logging functionality without mythlogserver is to compile with the --disable-logserver config option.

I'll start looking into that tomorrow. Anyone know offhand if the current 0.27 ebuilds use that?

Tom

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Jaglover · Posted: Fri Dec 06, 2013 12:26 am Post subject:

I'm too lazy to go and hook up a monitor to my backend (which is not in an easily accessible location). There is no trace of logging into any files.
In my frontend, however, I can see it logs to tty1 - the console I used to log in and run startx ...
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Yea, that would make sense. I logged an enhancement bug suggesting a possible USE flag to allow the --disable-logserver configure option:

https://bugs.gentoo.org/show_bug.cgi?id=493396

When I upgrade I'm going to add that option in an overlay ebuild.

Tom

depontius · Advocate Joined: 05 May 2004 Posts: 3509

Request for simple help...

I believe I've finally gotten past (avoiding) the gnome-4.8 upgrade of last week and it's getting time to upgrade to 0.26. I need to do the timezone table check, for real. At this point I've forgotten the link that told how to do it. At the MythTV wiki they have a few cryptic lines for 0.26 or 0.27, but I don't see what to run if you're checking for moving to 0.26. Nor are the lines complete, they assume that you've already started a mysql session.

I only do anything with mysql when I have mythtv problems - like right now, so I've forgotten how to get started. Does anyone have a "complete" procedure on how to check for the timezone table?
_________________
.sigs waste space and bandwidth

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

I've never seen those instructions, and I'm not sure what you mean by a "complete" check of the timezone tables. I work with MySQL all the time. The only way I know of to load the timezone tables is via the sql generated by the mysql_tzinfo_to_sql program, as described in the MythTV wikis. Assuming you've done that they should be fine.

As far as I know, the following (which I ran on my backend system) would not work without them (note that I've obfiscated the password):

depontius · Advocate Joined: 05 May 2004 Posts: 3509

Thanks, that's what I was looking for:

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

I'm testing things for my upgrade. This one REALLY had me going:

If you need/want to apply patches (for example user patches in /etc/portage/patches/www-apps/mythweb-0.27) for mythweb-0.27, the paths in the patch files need to have mythweb-0.27/ at the beginning....for example:

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Another note: In several places here I referred to the new 0.27 configure option --disable-logserver. That was incorrect...the correct option is --disable-mythlogserver.

Tom

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Cool! My upgrade from 0.25.3 to 0.27 appears to have gone flawlessly! Very cool stuff.

I'm going to note on that bug I opened that the addition of --disable-mythlogserver to the ebuild worked great. Everything is logging to /var/log/mythtv as expected with no mythlogserver running.

Tom

depontius · Advocate Joined: 05 May 2004 Posts: 3509

Are you sure that "--disable-mythlogserver" option is new to 0.27 - is there a chance it's on 0.26 as well?

I sometimes use the logs to find out when things go bad. I presume "rm -f" or "chmod -x" means that nothing will get logged? Is it a certainty that the logserver is going to misbehave? Does anyone understand the problem, or have fixes?
_________________
.sigs waste space and bandwidth

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

The --disable-mythlogserver option is only in 0.27 as far as I know.

I'm actually not clear as to specifically what issues people had with mythlogserver, but apparently some did. I get the impression that one of it's main purposes was for debugging, as you'd never loose the last log message from a program that aborts for example. For most peoples purposes, I think there's no real advantage of using it, which is why they made it optional in 0.27.

And yes, I think that is you remove the executable or it's execute permissions, you'll just get console logging.

I'm really happy with how my 0.25.3 -> 2.37 upgrade has gone...everything's working great. I copied the standard ebuild to my local overlay (with a sym link to the live files directory) and made this change to it:

depontius · Advocate Joined: 05 May 2004 Posts: 3509

Juat a note.... I moved to 0.26 Sunday, and so far (knock on wood) mythlogserver has stayed pretty much below 1% CPU utilization.
_________________
.sigs waste space and bandwidth

Jaglover · Posted: Tue Dec 10, 2013 9:07 pm Post subject:

Good luck, mine consumed twice as much as mythbackend process, and sometimes just filled the memory with numerous logserver instances, slowing the backend to a crawl. I had less trouble with 0.27, but the issue stiil is there.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly!

tld · Veteran Joined: 09 Dec 2003 Posts: 1816

Just a follow up FYI: As I noted in the gentoo bug I started around supporting --disable-mythlogserver, I ended up having to abandon that one.

Apparently, while that option causes the logs to be written in a manner much like that in 0.25.3, there apparently is no longer any logic to reopen log files via a HUP signal anywhere except in mythlogserver. As a result, after a weekly log rotation, I found my mythbackend actively logging to the rotated file (under the "old" directory with a .1 at the end of the file name).

So apparently, unless you can enable syslog logging, there's simply no way to handle log rotation with that option.

I just recompiled today with the standard ebuild. So far it doesn't appear that anything I do causes any issues with multiple mytlogserver processes etc.

Tom

ashdash · Posted: Sun Dec 29, 2013 8:43 pm Post subject: Be aware of this bug.

If you are going to emerge mythtv-0.27_p20130924 or mythtv-0.27_p20131124 be aware of this bug

https://bugs.gentoo.org/show_bug.cgi?id=496046 which is duplicate of 489912.

CFLAGS with spaces are getting mangled. I had to disable distcc and set CFLAGS="-O2 -pipe -march=native" from
CFLAGS="-O2 -pipe -march=amdfam10 -mcx16 -msahf -mno-movbe -mno-aes -mno-pclmul -mpopcnt -mabm -mno-lwp -mno-fma -mno-fma4 -mno-xop -mno-bmi -mno-tbm -mno-avx -mno-sse4.2 -mno-sse4.1 --param l1-cache-size=64 --param l1-cache-line-size=64 --param l2-cache-size=1024 -mtune=amdfam10".

All the flags after -pipe are derived from the command gcc -march=native -E -v - </dev/null 2>&1 | grep cc1 so that the correct cpu optimizations get passed to the other machines in the distcc pool. I guess, depending on the processors being used, some strings could be ok. My processors are AMD of various flavours.