Thanks! That's a very useful link actually. I've been trying to figure out what all the options are. For some reason it didn't catch my attention during my 'grep's. That fixes all the above problems, I can even use -u and -g now.
I created the caldavd user in PostgreSQL and created the caldavd database (I'm using -u caldavd -g caldavd again). I noticed this connects to the database using the caldavd user. Now I get this error:
2011-12-27 16:17:47-0500 [-] [txdav.common.datastore.util.UpgradeDatabaseSchemaService#warn] Beginning database schema check.
2011-12-27 16:17:47-0500 [-] [txdav.common.datastore.util.UpgradeDatabaseSchemaService#warn] Required schema version: 3.
2011-12-27 16:17:47-0500 [-] Unhandled error in Deferred:
2011-12-27 16:17:47-0500 [-] Unhandled Error
Traceback (most recent call last):
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 388, in errback
self._startRunCallbacks(fail)
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 455, in _startRunCallbacks
self._runCallbacks()
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 542, in _runCallbacks
current.result = callback(current.result, *args, **kw)
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 1076, in gotResult
_inlineCallbacks(r, g, deferred)
--- <exception caught here> ---
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 1018, in _inlineCallbacks
result = result.throwExceptionIntoGenerator(g)
File "/usr/lib/python2.7/dist-packages/twisted/python/failure.py", line 350, in throwExceptionIntoGenerator
return g.throw(self.type, self.value, self.tb)
File "/usr/local/lib/python2.7/dist-packages/txdav/common/datastore/util.py", line 271, in doUpgrade
actual_version = yield sqlTxn.schemaVersion()
File "/usr/lib/python2.7/dist-packages/twisted/internet/defer.py", line 1018, in _inlineCallbacks
result = result.throwExceptionIntoGenerator(g)
File "/usr/lib/python2.7/dist-packages/twisted/python/failure.py", line 350, in throwExceptionIntoGenerator
return g.throw(self.type, self.value, self.tb)
File "/usr/local/lib/python2.7/dist-packages/txdav/common/datastore/sql.py", line 233, in schemaVersion
result = yield self._schemaVersion.on(self)
File "/usr/local/lib/python2.7/dist-packages/twext/internet/threadutils.py", line 64, in _oneWorkUnit
result = instruction()
File "/usr/local/lib/python2.7/dist-packages/twext/enterprise/adbapi2.py", line 155, in <lambda>
lambda : self._reallyExecSQL(*args, **kw)
File "/usr/local/lib/python2.7/dist-packages/twext/enterprise/adbapi2.py", line 136, in _reallyExecSQL
result = self._reallyExecSQL(sql, args, raiseOnZeroRowCount)
File "/usr/local/lib/python2.7/dist-packages/twext/enterprise/adbapi2.py", line 130, in _reallyExecSQL
self._cursor.execute(sql, args)
File "/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/dbapiclient.py", line 76, in execute
self.realCursor.execute(sql, args)
File "/usr/local/lib/python2.7/dist-packages/pgdb.py", line 259, in execute
self.executemany(operation, (params,))
File "/usr/local/lib/python2.7/dist-packages/pgdb.py", line 289, in executemany
raise DatabaseError("error '%s' in '%s'" % (msg, sql))
pg.DatabaseError: error 'ERROR: relation "calendarserver" does not exist
LINE 1: select VALUE from CALENDARSERVER where NAME = 'VERSION'
^
' in 'select VALUE from CALENDARSERVER where NAME = 'VERSION''
I would have expected calendarserver to set up its own tables/schema (as it appears to be attempting to do). Can I force it to rebuild the database?
Thanks,
Jim
On Tue, Dec 27, 2011 at 1:40 PM, Andre LaBranche
<dre@apple.com> wrote:
Hi,
You will also need to specify DSN which tells calendar server how to connect to postgres.
You can also set ServerEnabled to false for memcached to keep calendar server from spawning memcached.
I can provide examples once I get back home to a real machine...
Cheers,
-dre
Sent from my iPhone
All,
I'm quite new to this calendarserver thing. I'm trying to set up 3.1 (from tags/releases in Subversion) on a 64-bit Ubuntu 11.10 server installation. It's been a disaster. I was able to run calendarserver from the SVN directory using ./run and then I "installed" it using ./run -i. That's where things went bad. Some issues I've worked around, but I'd like to get good solutions for. Currently I'm stuck trying to figure out what's going on with PostgreSQL. Perhaps you guys can clue me in on how to properly fix the following issues.
This is how I'm currently executing caldavd. Once it works I'll put it into init.d and run it as a proper server. The () are because I *was* running with those flags (see part 1) and have given up on them.
sudo caldavd -X -L -f <path to caldavd.plist file> (-u caldavd -g caldavd)
1. As I understand it, it is good practice to run Linux services that are exposed to the internet as a restricted user. Given that and existence of the -u and -g flags for caldavd, I figured those flags would be the right way to do things. That doesn't work at all.
The first problem is that caldavd wants to write to /var/run/ccs_postgresql_xxxxx which it can't do because /var belongs to root and it attempts to do this after the setuid() call to become the caldavd user. So, I went into caldavd.plist and changed the "DataRoot" to a different directory and gave it to caldavd:caldavd.
Then I tried again. This time it failed because it appears that calendarserver attempts to start postgresql whether or not the postgresql service is already running on the local box. To do this, it tries to setuid() again, this time to ROOT! If the entire point of setting a user for a service is to restrict it's privileges, why would one of those privs be to allow setuid to root? That makes no sense.
So I gave up, now I'm trying to run it as root to see if the 3.1 release even functions. How do I make this server run as somebody other than root?
2. Running as root (without the -u and -g flags shown above) gets the application started and then it fails attempting to start memcached. I get the error "cannot listen on tcp port". This is because memcachd is a system service and is *already running*. Again, why is calendarserver attempting to start a service that's already running?
So I stopped memcachd and tried again.
3. Calendarserver now successfully starts and connects to memcached but it cannot connect to postgresql. This I haven't yet worked around. I get the following error:
2011-12-27 12:43:11-0500 [-] [txdav.base.datastore.subpostgres#error] [Failure instance: Traceback: <class 'pg.InternalError'>: could not connect to server: No such file or directory
Is the server running locally and accepting
connections on Unix domain socket "/tmp/ccs_postgres_f5e263275ce7de576bd42ca08db4fc56/.s.PGSQL.5432"?
/usr/lib/python2.7/dist-packages/twisted/internet/_baseprocess.py:60:maybeCallProcessEnded
/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/subpostgres.py:92:processEnded
/usr/lib/python2.7/dist-packages/twisted/internet/defer.py:361:callback
/usr/lib/python2.7/dist-packages/twisted/internet/defer.py:455:_startRunCallbacks
--- <exception caught here> ---
/usr/lib/python2.7/dist-packages/twisted/internet/defer.py:542:_runCallbacks
/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/subpostgres.py:367:gotReady
/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/subpostgres.py:269:ready
/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/subpostgres.py:261:produceConnection
/usr/local/lib/python2.7/dist-packages/txdav/base/datastore/dbapiclient.py:196:connect
/usr/local/lib/python2.7/dist-packages/pgdb.py:482:connect
]
Here I'm stuck. Quite clearly calendarserver is expecting to launch the postgresql service (which, again, doesn't make sense to me), but then it can't connect when it does. Well, at least I'm assuming calendarserver successfully launches postgresql because I don't get the setuid() error I got in part 1 when it tries to setuid() to root in order to launch postgresql. I can't figure out how to fix this. Any thoughts?
Thanks for the help,
Jim