Main > Forum/Website Discussion

*** SERVER ISSUES *** - we are currently experiencing a heavy load on the server

<< < (10/14) > >>

Phreakwars:
Are there other domains on the server that you are sharing, or is it a dedicated server? You can experience problems of the like when you are sharing with several domains. That's just kind of the nature of it when it comes to some of these storage space companies. You could try asking to be switched to a different rack that isn't being utilized as much by others.

saint:

--- Quote from: Phreakwars on May 11, 2016, 03:23:21 pm ---Are there other domains on the server that you are sharing, or is it a dedicated server? You can experience problems of the like when you are sharing with several domains. That's just kind of the nature of it when it comes to some of these storage space companies. You could try asking to be switched to a different rack that isn't being utilized as much by others.

--- End quote ---

Dedicated server.

shock_:
Saint that does sound encouraging. I thought maybe there was some manual breakfast service restarts going on, that does fit the pattern of downtime :)

One other really good thing to do is install New Relic PHP extensions. Not too hard to get going, and the application monitoring will show you exactly what is going on over time, and allow for some deep diving into slow transactions for the first couple of weeks while they give you the full, expensive report package. Their free tier is still worth leaving up though too, once the trial expires. They also have a server health metrics thing that will graph things like your iotop and more.

The reporting in New Relic shouldn't have any personal info exposed and if you later feel like more public discussion is useful, well, we all love graphs :)

Malenko:
Install atop

Run it. Press "i" to set the interval, I like 3 seconds.

See if the disks are actually busy and what the request latency is.  See if there's a bunch of CPU iowait. Are the CPUs busy? Is the box actively swapping to disk?

If the disk latency is higher than low single digit milliseconds but the disks don't look busy, check the SMART stats with smartctl. Look for a high number of reallocated sectors. This would indicate the drive is failing.

If all this looks okay and the drives are mostly idle, the box isn't swapping, the CPUs aren't busy and there's no high iowait, look at mysql.

Enable the slow query log: set global slow_query_log=on;
See where mysql is writing the slow log: show variables; look for slow_query_log_location or similar.
Find a query or queries that are logged slow, and run explain on it in mysql: explain <paste query here>;
Look for things that aren't using indexes, anything that says "null null all" is doing a full table scan and is going to be hella slow.

Let me know how all this crap looks and we'll go from there.

Phreakwars:
Would also like to add, try asking one of the guys on The Admin Zone.

https://theadminzone.com/

Been a member there for years, they have some of the best guys who can also help solve the issue. You'll be able to find people who know the in's and out's of every aspect of running a forum and also know all about the SMF software. The SMF developers themselve's also frequent the forum.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version