fdr

25 ms·

211.
by fdr 12y ago
You are absolutely right. That's because when a parametric query arrives, the parameters are unbound and the planner cannot take advantage of fit-to-purpose selectivity estimation. It must instead estimate generically. Newish versions
212.
by fdr 12y ago
Only because the idea of manipulating clients of the database to do HA is somewhat heretical.
213.
by fdr 12y ago
I'd say the way failover/HA is done at Heroku is straightforward. Will (the designer) or I plan to write about it some day, pending laziness. It took some time to figure out because it required breaking some orthodoxy, but I'
214.
by fdr 12y ago
I think you may be interested in the "h-index" and related metrics. http://www.cs.ucla.edu/~palsberg/h-number.html http://en.wikipedia.org/wiki/H-index http://researchguides.u
215.
by fdr 13y ago
That's very humbling to hear (I've spent, cumulatively, a lot of time maintaining WAL-E), yet log shipping is a feature of many other database systems. That said, I do write and maintain WAL-E for reasons that go beyond mere vanit
216.
by fdr 13y ago
> No, I didn't take storage or IOPs into account but you do get more memory and the flexibility if the Heroku options don't fit you. No doubt about that. To date, Heroku's product model has preference for fewer, but common
217.
by fdr 13y ago
> From a very quick glance at the small end the Amazon options are a little more than half the price (Ireland single AZ instance pricing and even less reserved) but multi-AZ options are probably very similar. Did you add disk? People se
218.
by fdr 13y ago
Thanks for the idea. A modified variant created https://github.com/heroku/toolbelt/pull/74
219.
by fdr 13y ago
You are right: in the fast case of "NULL" the amount of physical work amounts to diddling catalogs around rather than copying potentially gigabytes of data. So in any case a lock is taken, but I've never seen anyone get too b
220.
by fdr 13y ago
I'll forward your remarks to the devcenter staff and the steward of the changes; I'm sure both parties will appreciate it a lot.
221.
by fdr 13y ago
I don't think worrying about database portability often makes much sense for projects with beyond the simplest access patterns and projects that one does not ship to a client, where one anticipates needing to support many possible data
222.
by fdr 13y ago
Meh. Many well-established companies in a lick of trouble run negative P/E ratios for years...as in, the firm is losing money. There may be arguments for or against a given share price, but anything coming strictly from very high P/E va
223.
by fdr 13y ago
An immutable mark will allow the optimizer to constant-fold an expression that contains only IMMUTABLE operations. It basically has to believe you on this one, and being wrong is not good at all. STRICT is less advisory: if specified, the
224.
by fdr 13y ago
Database constraints usually catch these problems in event of re-submission, especially if the client can assign primary keys (e.g., a UUIDv4) a-priori, but this also tends to be true in simpler cases, too. All in all, I am not sure if any
225.
by fdr 13y ago
I don't like to strenuously object anything in general, but I'll make an exception this time: this advice makes me burst a vessel every time I hear it, because I must confess that the number of code bases I've seen that are chronically over
226.
by fdr 13y ago
Wouldn't that be nice! I work on Heroku, on the databases, which are rather thorny in that department... ...it's going to take some evolution in how applications are written to make that work not uglifically (and it'll probably be ugly at f
227.
by fdr 14y ago
Heroku offers one database management system right now: Postgres, and its principal replication technology is not based on statement-based replication, and largely for this reason. Using something like PGPool can make the problem come righ
228.
by fdr 14y ago
I am a mere small fry employee at Heroku. I nearly exclusively work on the efforts concerning Postgres. If you stopped using Heroku's product because you thought it would not pragmatically meet your needs: I'm all for that. However, by des
229.
by fdr 14y ago
I think most people used to a materialized view features in proprietary databases will find the incarnation of Materialized Views as committed quite limiting...nevertheless, it seems important to me that this foundational work has been inte
230.
by fdr 14y ago
I would like to second this. It's imperfect, but a lot of the fiddly bits -- hot plugging monitors, web cams, et al -- work, and the bug reporting and fixing is tracked: https://bugs.launchpad.net/dell-sputnik That's also a reasonably go
231.
by fdr 14y ago
> An more detailed writeup, but hardly satisfying. Like most of Heroku's technical choices made in the name of "simplicity" and "robustness." > But I guess I should accept that Heroku will never be a technically impressive PaaS prov
232.
by fdr 14y ago
Shucks. Well, I'll try to not let you down in the coming year, insomuch as I can muster.
233.
by fdr 14y ago
I think you petered at at the end of my wall of text ;) But seriously, I'm glad that I was not totally off base in letting my mind drift to IPv6. Thanks.
234.
by fdr 14y ago
Well, I'll tell you why it is not implemented that way. I hope the restriction can be lifted some day. I am a member of the Heroku staff related to the matters at hand. The problem is the sheer number of Heroku Runtime machines which are
235.
by fdr 14y ago
False (I work there), although some of my colleagues have submitted patches. Commit access is handed out very selectively at postgresql.org, and committing patches is not the bottleneck in making features (reviewing and writing the patches
236.
by fdr 14y ago
Naah, I wouldn't go so far to say it's a good thing that it's missing, but it's hellaciously hard to implement well. I think the sentiment is rendered well in this mail, which took place after extensive review and hacking. In it, Robert
237.
by fdr 14y ago
The indexes, though, both benefit (locality) and suffer from (lock contention) such random identifiers. That having been said, many data sets work with UUID just fine and the fact that such sets are easier to catenate and compare is very co
238.
by fdr 14y ago
Solution: remove all the error messages for common misspellings of what people want to do, or alternatively never deprecate any user interaction choice ever. One may make the argument that removing the default behind 'heroku console' was to
239.
by fdr 14y ago
The downside is use of more memory that can cause an out-of-memory condition, and the subsequent guesswork as to what the concurrency factor ought to be. However, it is often a really good idea for most applications that don't need a wh
240.
by fdr 14y ago
I have been using the current-generation XPS-13 (bought with Ubuntu pre-installed) and have not had hard lockups. I did have one kind of bug that seemed to make X completely lose its mind, but the virtual terminals still worked. I have upg

More ›