At 02:37 PM 7/15/2012 -0400, Jared Camins-Esakov wrote:
Using 3.6.1 in production (no longer on sandbox, we're looking at 3.8)
modifying frameworks never gave "instant gratification" due to memcached,
so I use netcat:
echo -e "flush_all\nquit\n" | nc localhost 11211
(and even have a temp 1 min cron job if I need more than a very simple
change.)
If the only thing using memcached is Koha, you might find it easier to
simply have a cron job that runs `service memcached restart` every hour.
Koha is only about 10% of our memcached load, but I do cron "flush-all",
"sleep 10" then "restart" every 24 hrs (at 3am, low/no use, gets around the
time delay "gotcha".) Also, as far as I know "restart" might not totally
*flush* memcached (there's a lot of debate about it, particularly in the
php world where memcached's garbage collection seems to be a dirty word)
but I haven't done any serious testing. At least "flush-all" is virtually
instantaneous, and meets the needs of Koha frameworks.
Do not mix and match files from different versions. My recommendation
would be to upgrade to a recent version of 3.6.x.
Thanks for the confirmation. That was my gut feeling. But ... I really am
trying to avoid trashing the work I've done on 3.8 on the sandbox (which is
too old to successfully run two versions of Koha.) I'm nowhere near ready
at this time to put 3.8 into production, so I guess my $64,000 question is
"what is the most stable version of 3.6 that I can upgrade the production
server to, with the least chance of running into problems?" There's
probably a lot of other bug fixes and enhancements that could be
useful. Maybe wait for your 3.6.7? :=)
[snip] (WARNING: the following is irrelevant to your question) I would
argue this is a bug (reported it as one, in fact), but that's the way it
is, so -- for the moment -- no need to mark both the tag and subfield
mandatory. Based on the way the MARC standard is written, marking a
subfield mandatory should make it so that if any subfield in that tag is
populated, that subfield must be populated, but not require that the tag
be populated at all. However, that's not the way it works, so it's kind of
a moot point.
Sorry, but I can't find the bug you reported (number?) From my
perspective, it was a [probaly very rare] combination of circumstances and
as you say "a moot point", hence my calling it an "oddity" rather than a
bug. We're back on track, the cataloguers see a red asterix, and the help
file (ours, not LoC) pops up when requested.
Thanks again and regards,
Paul
_______________________________________________
Koha-devel mailing list
[email protected]
http://lists.koha-community.org/cgi-bin/mailman/listinfo/koha-devel
website : http://www.koha-community.org/
git : http://git.koha-community.org/
bugs : http://bugs.koha-community.org/
