Is this reproducible? If so, instead of timing
t0=time.time()
orders = db(db.orders.id > 0).select()
logging.info('time to fetch %s' % (time.time()-t0))
try time
sql = db(db.orders.id > 0)._select()
t0=time.time()
db.executesql(sql)
logging.info('time to fetch %s' % (time.time()-t0))
orders=[]
I do not think this time is spend in web2py. I think - for some reason
- it is spent communication with mysql.
On Dec 25, 4:25 am, Kenneth Lundström <[email protected]>
wrote:
> 980 rows, 10 columns of data,
> 38 sec.
>
> MySQL server is on the same server, Apache server is not local, but the
> delay is allways 38-39 seconds.
>
> Kenneth
>
> > No idea. How much data.
> > What db? is it local? Probably something is wrong anyway.
>
> > On Dec 24, 5:55 pm, Kenneth Lundstr m<[email protected]>
> > wrote:
> >> I finally had time to find out a little about loggin.conf.
>
> >> In my logs/web2py.log I get this line:
> >> 2010-12-25 01:42:30,097 - root - INFO - time to fetch 1.81220602989
>
> >> So it takes under 2 seconds to fetch the data and the rest 37 to show
> >> it? How come?
>
> >> Kenneth
>
> >>> Were is this logged, I could not find anything in any logs?
> >>> Kenneth
> >>>> try
> >>>> import time, logging
> >>>> def testing():
> >>>> t0=time.time()
> >>>> orders = db(db.orders.id> 0).select()
> >>>> logging.info('time to fetch %s' % (time.time()-t0))
> >>>> return orders
> >>>> so you can isolate the problem and see if it is in fetching or
> >>>> somewhere else (for example session locking).
> >>>> On Dec 2, 4:21 pm, Kenneth Lundstr m<[email protected]>
> >>>> wrote:
> >>>>>> When you test is that the only active connection to the db?
> >>>>> I tested it on an application running on the test instance. There could
> >>>>> have been lite activity one production site.
> >>>>> But when I tested loading the same function five times in about 10
> >>>>> minutes I allways 39 seconds to load the data.
> >>>>> Kenneth
> >>>>>> 2010/12/2 Kenneth Lundstr m<[email protected]>:
> >>>>>>>> Please tell us more about the setup. Are the three instances behind
> >>>>>>>> running on the same server?
> >>>>>>> Yes.
> >>>>>>>> why three? What do they do?
> >>>>>>> One is production, one is testing and last one is development. I
> >>>>>>> had two
> >>>>>>> instances running on a virtual server before and it worked fine.
> >>>>>>>> This is not normal but I suspect the problem is with database.
> >>>>>>>> Did you set DAL(...,pool_size=10)?
> >>>>>>> No, but I tried but no change.
> >>>>>>> If I try to select more rows:
> >>>>>>> def testing():
> >>>>>>> orders = db(db.orders.id> 0).select()
> >>>>>>> return orders
> >>>>>>> It takes 80 seconds to return 1608 rows.
> >>>>>>> If I change the return to return len(troops) it only takes 1,5
> >>>>>>> seconds.
> >>>>>>> I have no view defined, just trying out the database.
> >>>>>>> Kenneth
> >>>>>>>> On Dec 2, 5:13 am, Kenneth Lundstr m<[email protected]>
> >>>>>>>> wrote:
> >>>>>>>>> Hello list,
> >>>>>>>>> need some help finding out why my server is so slow.
> >>>>>>>>> The server is with a Intel Celeron 2.66 GHz CPU, 4 GB of memory
> >>>>>>>>> CentOS 5.5 64-bit, Apache 2.2.3, MySQL 5.0.77, mod_wsgi, Web2py
> >>>>>>>>> 1.89.5
> >>>>>>>>> The server is dedicated to web2py, there is three instances of
> >>>>>>>>> web2py
> >>>>>>>>> running.
> >>>>>>>>> I have migrate=False, sessions on disc.
> >>>>>>>>> If I try the following code
> >>>>>>>>> def testing():
> >>>>>>>>> customers = db(db.customer.id> 0).select()
> >>>>>>>>> return customers
> >>>>>>>>> takes 39 seconds, 5 times in a row, to return 979 rows.
> >>>>>>>>> If I put db.customer< 100 it takes 4,5 seconds to return 87
> >>>>>>>>> rows.
> >>>>>>>>> Is it just me or are this a bit long times?
> >>>>>>>>> Kenneth
>
>
