| 14:30.32 | *** join/#udbug unum (n=unum@li32-23.members.linode.com) |
| 14:37.48 | *** join/#udbug Yorokobi (n=Yorokobi@unaffiliated/yorokobi) |
| 14:37.49 | *** mode/#udbug [+o Yorokobi] by ChanServ |
| 16:22.47 | *** join/#udbug whiteinge (n=whiteing@oalug/member/whiteinge) |
| 18:48.47 | *** join/#udbug SunSparc (n=sparky@173-8-91-177-Utah.hfc.comcastbusiness.net) |
| 18:48.48 | *** mode/#udbug [+o SunSparc] by ChanServ |
| 21:58.18 | *** join/#udbug whiteinge (n=whiteing@whiteinge.dsl.xmission.com) |
| 22:03.43 | *** join/#udbug whiteinge (n=whiteing@whiteinge.dsl.xmission.com) |
| 22:12.24 | *** join/#udbug eggyknap (n=jtolley@unaffiliated/eggyknap) |
| 22:12.24 | *** join/#udbug pelaofeliz (n=PelaoFel@67.108.236.230) |
| 22:12.24 | *** join/#udbug DexterTheDragon (n=admin@oalug/member/blackchaos) |
| 22:12.24 | *** join/#udbug ChanServ (ChanServ@services.) |
| 22:12.24 | *** mode/#udbug [+ooo eggyknap DexterTheDragon ChanServ] by irc.freenode.net |
| 22:27.36 | whiteinge | Postgres Q: Given a simple table used to store business-card info, how big could it grow before schema changes would start to take a prohibitively long time (assuming a moderately loaded production server)? |
| 22:27.46 | whiteinge | 200 million? |
| 22:27.48 | whiteinge | more? |
| 22:28.03 | whiteinge | ...or did I just demonstrate a profound lack of database understanding? |
| 22:31.34 | Yorokobi | Depends on # of writes to the DB, number of reads, how often the DB is vacuum'd |
| 22:57.56 | *** join/#udbug Yorokobi (n=Yorokobi@unaffiliated/yorokobi) |
| 22:57.56 | *** mode/#udbug [+o Yorokobi] by ChanServ |
| 23:16.52 | whiteinge | Say 20% writes, 80% reads, and vacuum-ed semi-often. Is that enough info to start making educated guesses, or is it better to just wait and see? |
| 23:17.05 | whiteinge | I'm in the process of replacing an old, badly-designed system and I'm wondering if I should just use a NoSQL database for the business-cards table. |
| 23:18.27 | whiteinge | NoSQL seems particularly well-suited to this kind of data since it tends to be isolated and the schema changes frequently (add a work fax, now we need a second cell number field, etc.) |
| 23:19.32 | whiteinge | But, of course, it's much more work to set up and query, so I'd just as soon stick it all in Postgres if it isn't going to cause any headaches every time the schema changes. |