As you may know I have spent some time recently working on the hyperkitty program. The idea being to offer a newinterface to the archives in mailman 3 (which has never been closer to arelease).
Hyperkitty aims at implementing a numbers of the ideas developed by MáirínDuffy in her blog posts. The main one being to try to unify mailing lists andforum (ie: providing a web-interface to mailing lists).
In this quest, I have started to look some time ago to MongoDB. It is a NoSQL database which is becoming quitepopular (probably helped with its integration into openshift). The results were satisfying but then a bigquestion came up:
· Do we really want to impose the burden of 2 different database systems toour sysadmin for a mailman archives interface ?
Of course, if we can avoid it, we wouldprefer to do it.
But MongoDB was performing reallynicely with regards to searching the archives. So testing was needed.
[size=11.5pt]The machine on which I ran the test has:
[size=10.0pt]· [size=11.5pt]16G of ram
[size=10.0pt]· [size=11.5pt]4 cores
[size=10.0pt]· [size=11.5pt]2x1To in RAID1[size=11.5pt]All this operated by RHEL 6.2 (Santiago).
[size=10.0pt]· [size=11.5pt]PostgreSQL version 8.4.9
[size=10.0pt]· [size=11.5pt]MongoDB version 1.8.2
[size=11.5pt]I used the archives from the[size=11.5pt] [size=11.5pt]devel[size=11.5pt] [size=11.5pt]mailing list, since its creation in 2002.
[size=10.0pt]· [size=11.5pt]PostgreSQL loaded 166672 emails
[size=10.0pt]· [size=11.5pt]MongoDB loaded 166642 emails
[size=11.5pt]So there is a difference of 30 emails which I considered to be negligible for the tests.
来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/301743/viewspace-731259/，如需转载，请注明出处，否则将追究法律责任。