Previously we only used these caches for Tornado requests, because we
were not updating memcached when e.g. the user's pointer changed, and
so functions like update_pointer would not work correctly.
Now that we are updated memcached when the User and UserProfile
objects change, we can use these for all requests.
This saves 2 database queries on every Django request to the server.
(imported from commit aa5bffd885d14bde38b95e80a226bd5ab66f253d)
This should substantially decrease the amount of server load generated
by the userpresence system.
I tested that this indeed was indeed saving one query on
/json/update_active_status requests on my laptop with 2 users from the
humbughq.com realm logged in.
(imported from commit 03e9d4eb95b9f664d489862684ae162db2076e08)
This cache filling code takes about 5 seconds to run, which means it
will finish before Tornado starts reading from this cache, but if that
were to change, it would be much better to have made at least some
progress.
(imported from commit 60a3420cdb9ddf331d83573a3fdb6be1a5ee5a4f)
Previously we were calling select_related() on Message.client, which
doesn't exist. It seems kinda poor that this doesn't raise an
exception.
I believe this issue was causing us to do very large numbers of
database queries during get_updates calls during server restarts.
(imported from commit b79bd698820fbd9dd82bd61fc175c32cd5ce6d05)
Since we flush memcached when we do a server restart, the flurry of
get_updates requests that fly in afterwards are all cache misses for
getting the User/UserProfile objects, so Tornado ends up spending
around 70ms per get_updates request rather than the usual 1-2ms.
So this should substantially improve our Tornado performance around
server restarts.
(imported from commit 07b8126bdfd4ff14e4c3362f9eda1fe5fd571c5b)
Sometimes Dropbox shares with /s/ and sometimes with /sh/,
and I'm not sure which controls it, but we should deal with both.
(imported from commit 2222450f25c418b5fbd60ab2c30477467e34c0d1)
This avoids our repeatedly retrying to fetch a tweet that doesn't
exist from the Twitter API.
(imported from commit b4ca1060d03da21e7e59e5b99e682d2e8457df15)
This should substantially improve the repeat-rendering time for pages
with large numbers of tweets since we don't need to go all the way to
twitter.com, which can take like a second, to render tweets properly.
To deploy this commit properly, one needs to run
./manage.py createcachetable third_party_api_results
(imported from commit 01b528e61f9dde2ee718bdec0490088907b6017e)
This commit adds a dependency on python-twitter,
whose upstream is at https://github.com/bear/python-twitter,
and which for now needs to manually be installed on our
servers from the Debian package in sid.
(imported from commit 80cd9f4f59a6f0de6b75ac95e412c69e2a2e2490)
This uses the unauthed v1 of the Twitter API, which is going to go
away soon, but it's fine as an interim measure.
(imported from commit 709a250271321f5479854a363875c9da43e6382d)
Prior to this change, any stream message sent by internal_send_message
could only be in the realm of the sender.
This was a problem most notably for... the tutorial bot, with the
hilarious consequence that the tutorial worked fine in humbughq.com,
but failed to start anywhere else.
(imported from commit 33a904a28e3a57e1a2cf9172c2e2a75b50967a50)
With this change,
pkill -SIGUSR1 -f runtornado
will dump the stack and SIGUSR2 will enable an interactive debugging
session.
This fixes#613 for Tornado which was the original motive for that
ticket; I'm not sure whether we want to do this for our Django
processes as well, but it would be easy to do so if we did.
(imported from commit a7de7c6070f4bf0404bed6f434e6a6b291d66a26)
Even though they look like images, they're not -- you need to
append ?dl=1 to get the image version.
(imported from commit 2a05e7c58f475c908687110d9191f8709425c660)
At Ksplice we used /usr/bin/python because we shipped dependencies as Debian /
Red Hat packages, which would be installed against the system Python. We were
also very careful to use only Python 2.3 features so that even old system
Python would still work.
None of that is true at Humbug. We expect users to install dependencies
themselves, so it's more likely that the Python in $PATH is correct. On OS X
in particular, it's common to have five broken Python installs and there's no
expectation that /usr/bin/python is the right one.
The files which aren't marked executable are not interesting to run as scripts,
so we just remove the line there. (In general it's common to have libraries
that can also be executed, to run test cases or whatever, but that's not the
case here.)
(imported from commit 437d4aee2c6e66601ad3334eefd50749cce2eca6)
Previously we were doing 100 queries for new messages being sent to
the main hacker school channels; they were faster than many similar
instances because they were all done within 1 transaction, but still,
send_message_backend would be spending up to 200ms (and 148 queries)
querying the database with the previous code on prod; this new version
should do a fixed number of database queries per message.
(imported from commit 3799e63aebb6f017932ddb0fe1f6209281c0ddcf)
To work around the issue we're having with queue draining between
parallel blocking connections, use the same rabbitmq queue for both
activity and presence events, keyed on a 'type' flag in the message
itself.
(imported from commit 188e8fda1695734e52c5740db2195072cfc81479)
