1 |
kashani wrote: |
2 |
> Chris Bainbridge wrote: |
3 |
>> I'm curious what the problem is with bugzilla and it's db |
4 |
>> interactions? You're suggesting a specific issue rather than general |
5 |
>> db performance issues like fs, io scheduling, raid1, hyperthreads, |
6 |
>> etc.? |
7 |
> |
8 |
> It's most likely related to Bugzilla using MyISAM tables by default |
9 |
> which is fine in a small environment. However as concurrency grows along |
10 |
> with table size those full table locks begin to cause issues. |
11 |
> Additionally most www-apps are not known for their well thought out db |
12 |
> schema. Performance of the underlying hardware plays a part, but can be |
13 |
> overshadowed pretty quickly by query and table inefficiencies. |
14 |
> |
15 |
> The usual fix without touching the internals of the software is |
16 |
> doing master/slave replication and allowing the selects to happen on the |
17 |
> slave and writes on the master. However you would need to change most of |
18 |
> the queries to point to the slave server which is usually not trivial. |
19 |
> It sounds like this is the path the Infra team is pursuing. |
20 |
|
21 |
1++ |
22 |
|
23 |
You got it right :-) |
24 |
|
25 |
-- |
26 |
Lance Albertson <ramereth@g.o> |
27 |
Gentoo Infrastructure | Operations Manager |
28 |
|
29 |
--- |
30 |
GPG Public Key: <http://www.ramereth.net/lance.asc> |
31 |
Key fingerprint: 0423 92F3 544A 1282 5AB1 4D07 416F A15D 27F4 B742 |
32 |
|
33 |
ramereth/irc.freenode.net |