https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #14 from Sean Pringle ---
Ah, sorry. Dropped the ball.
Setting up a real test of both options on enwiki.eternallinks now.
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bu
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #13 from Greg Grossmeier ---
(suggestion: add it (the question for Sean) to the Scrum of Scrums)
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
___
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #12 from Andre Klapper ---
(In reply to Brad Jorsch from comment #9)
> Sean, how does my proposal in comment 7 sound to you?
Looks like we can't make Sean answer here. :(
--
You are receiving this mail because:
You are the assign
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #11 from Andre Klapper ---
Sean: How does anomie's proposal in comment 7 sound?
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #10 from Andre Klapper ---
Sean: How does anomie's proposal in comment 7 sound?
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #9 from Brad Jorsch ---
Sean, how does my proposal in comment 7 sound to you?
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
__
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #8 from Aaron Schulz ---
I'm OK with A too. The above approach looks reasonable. Indeed we will need
both extra columns with indexes and not just more prefixes since even FORCE is
not enough to get MySQL to be aware of the possible
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #7 from Brad Jorsch ---
I thought I had posted an analysis of these API modules somewhere, but now I
can't find it.
As noted already, the query for ApiQueryExtLinksUsage looks something like
this:
SELECT page_id,page_namespace,pa
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #6 from Sean Pringle ---
If the B method of paging works for the users, great. Let's see what Anomie
thinks.
--
You are receiving this mail because:
You are the assignee for the bug.
You are on the CC list for the bug.
___
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #5 from Aaron Schulz ---
With B, the API would would get the requested number of rows within a shard
(moving to the next one as needed to respect the limit). The API continue
parameter would include the shard and next link. The shar
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #4 from Sean Pringle ---
Option A means using FORCE INDEX I guess. Since multiple indexes would increase
the overall footprint it would need testing to see if there is a net
improvement. My guess would be probably not much.
Option
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
Aaron Schulz changed:
What|Removed |Added
CC||aschulz4...@gmail.com
--- Comment #3 fr
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
--- Comment #2 from Sean Pringle ---
enwiki> SELECT COUNT(*)
-> FROM externallinks
-> WHERE el_index LIKE 'http://gov.nih.nlm.ncbi.%';
+--+
| COUNT(*) |
+--+
| 951774 |
+--+
1 row in set (3 min 57.90 sec)
T
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
Andre Klapper changed:
What|Removed |Added
Keywords||performance
Priority|Unprior
https://bugzilla.wikimedia.org/show_bug.cgi?id=57176
Umherirrender changed:
What|Removed |Added
CC||umherirrender_de...@web.de
--- Comment
15 matches
Mail list logo
