New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
i could do that since begining but want to read some options.
thanks to all.
@Dedicados,
PM me if you want, I can work something out for you
@dedicados
+1 to optimization. Having had a look at the pasta - <2GB of RAM and 32bit system right now? Also running an older MySQL version?
SSD or SSD Cache plans would help, but might not be needed. I don't think you need a dedicated server either.
We do offer server management and tuning if needed as extra options
Do you use apc memcache varnish or nginx
Make a ioping and dd test for disks
Put a link to see this sites
Who is your hoster
Good point that has been overlooked so far.
108 queries per second and you're looking for hardware like that? Go with @prometeus VPS and use their MySQL offloading. Or BuyVM if you need one in the U.S.
@tsanten
./ioping -R /
--- / (simfs /vz/private/1318) ioping statistics ---
22371 requests completed in 3000.0 ms, 15583 iops, 60.9 mb/s
min/avg/max/mdev = 0.0/0.1/34.7/0.3 ms
root@pasta [~/ioping-0.6]#
@dedicados: If your servers are close to Italy, i'd say use Prometeus's MySQL server
@Kairus those where in 7 min, let me leave mysql work for 1 hour and i paste queries per second.
i use prometeus mysql offloading and i crash their server literally, they can confirm you that.
Do you ever done a
mysqlcheck -u root -pxxxxxxxx --auto-repair --optimize --all-databases
xxxxxxxx stands for your root mysql password
100q/s isn't a whole lot and we have a few people that blast well over that during their crons/etc.
-03:00:26- addons01.frantech.ca:/ :: root # mysqladmin status
Uptime: 2204668 Threads: 623 Questions: 3368448569 Slow queries: 31949 Opens: 3120561 Flush tables: 1 Open tables: 8192 Queries per second avg: 1527.871
Francisco
Indeed, he has very large tables, but after optimizations now runs with 6 cores without load going up to 20+
It looks like we are on the right track after a day of looking into it.
@tsanten yes i use it on the morning but this command: mysqlcheck -o -A
looks like is doing the same.
@Francisco yes, your sql has 25 days running, let me leave mine at least 1 day (if load dont make it to restart) and i show the info for "mysqladmin status"
Check the queries sec, not just the time up.
Francisco
Yes i see @Francisco, so many queries so something where bad,
but! now @prometeus fixed it, till now, load is no more 2, excellent support =D
dedicados,
@prometeus is a good choice, otherwise, get a dedi and have every thing set up by your own. I am not using the pl script, sql optimization is very different from one to another, mostly depends on what "kind of intensive" to due with.
Well, too bad it came after the refund
But Uncle Sal is a miracle worker, you are free to return
CloudVPS.com in Amsterdam 65 euros for 4GB ram, 80 GB HD, 8 cores. They have excellent reviews on webhosters.nl and ispgids.com.
Versio.nl's Cloud VPS, also in Amsterdam, is the most scalable. Up to 26 cores. 80 euros for 8GB ram, 800 GB HD, 8 cores, 16TB bandwidth, 8 IPv4 ips . 119 euros for 12 cores, 12GB ram, 1.2 TB HD, 24 TB bandwidth, 12 IPv4 ips. They also have good reviews on webhosters/ispgids but their website is in Dutch
thanks Maounique, of cours i still on prometeus.
Considering the high cost of SSD storage right now, I would say $50 per month is not a Reasonable budget considering you also require a decent amount of CPU cores and RAM on top of the pure SSD storage. $100 sounds more reasonable. Let us know who you end up signing up with.
@N_Serian i was with @prometeus , and now with help of all here, i was reading and finding how to optimize SQL, @prometeus did the last importan part, optimizing the TABLE of the database who need more power.
now load is 2 and 2.5MAX, been running like that for 4 hours
right now: Load Averages 1.01 1.04 1.08
Most of the problem where caused by two queries running against hundreds of thousands of records without the right index, so a full scan was perfomed at each request.
Adding the right index helped the optimizer to go straight to the result set.
The slow query log and some work with explain can help a lot.
I like this. I like this a lot.