I cannot follow your love for mongo, i have been forced to move away from it.
Let me list a few issues i have.
1. overloads my harddisks when i do bulk insert (completely kills everything else on the server) 2. cannot give space back to the OS, i bought fast disks, not big ones, i can delete old data if i need space (well, apparently not) 3. Memory consumption, i got 48gb mem, still this doesn't seem like enough for 50mio entries (these entries are very small, and not a lot of indexes either)
MongoEngine
1. The developer seems to actively commit on latest release, i did complain about this on the mailinglist 2. For some reason unknown, the mapping between pymongo and mongoengine isn't direct. This means that uuids will be screwed if you load and save them using mongoengine
Anyways, this is where i ditched mongo and moved along. It is important to keep in mind that using a new db isn't just glitter and good times, there's always new problems (yet nobody mentions these)
Anders, I thank you deeply for your couragous comment here - we need more people like you, that do not blindly advertise new technologies but seriously test them and report about their conclusions! I really fear a serious degradation in quality of the open source community in general, because all these conferences and events are more and more becoming (rouge) marketing events. There are absolutely no clear lines between sponsored marketing blah and scientific, unbiased reports. Criticism is becoming generally condemned as trolling and many speakers do not seem to have a problem with presenting themselves as blind followers, - aka "fanboys" - of technologies without even asking if there might be (new) problems - an attitude that might lead to discredit in the scientific world. (Not speakig specific about Peter here, I was not attending his talk, but to be meant as an observation from various different events I have visited before.) Scientific methodologies and priniciples like neutrality and scepticism should be seen as basic companions of open source software, not as hostile intrusion into a buddy-patterned social collective.
Comment
Hi
I cannot follow your love for mongo, i have been forced to move away from it.
Let me list a few issues i have.
1. overloads my harddisks when i do bulk insert (completely kills everything else on the server)
2. cannot give space back to the OS, i bought fast disks, not big ones, i can delete old data if i need space (well, apparently not)
3. Memory consumption, i got 48gb mem, still this doesn't seem like enough for 50mio entries (these entries are very small, and not a lot of indexes either)
MongoEngine
1. The developer seems to actively commit on latest release, i did complain about this on the mailinglist
2. For some reason unknown, the mapping between pymongo and mongoengine isn't direct. This means that uuids will be screwed if you load and save them using mongoengine
Anyways, this is where i ditched mongo and moved along. It is important to keep in mind that using a new db isn't just glitter and good times, there's always new problems (yet nobody mentions these)
Replies
Anders, I thank you deeply for your couragous comment here - we need more people like you, that do not blindly advertise new technologies but seriously test them and report about their conclusions!
I really fear a serious degradation in quality of the open source community in general, because all these conferences and events are more and more becoming (rouge) marketing events. There are absolutely no clear lines between sponsored marketing blah and scientific, unbiased reports. Criticism is becoming generally condemned as trolling and many speakers do not seem to have a problem with presenting themselves as blind followers, - aka "fanboys" - of technologies without even asking if there might be (new) problems - an attitude that might lead to discredit in the scientific world. (Not speakig specific about Peter here, I was not attending his talk, but to be meant as an observation from various different events I have visited before.)
Scientific methodologies and priniciples like neutrality and scepticism should be seen as basic companions of open source software, not as hostile intrusion into a buddy-patterned social collective.