SQL related server crash - at the end of my tether

[culprit]

Hi all,

I realise this problem encapsulates more than just SQL issues but its the closest topic I could find.

I'm running PHP / MySQL on an Apache server and it has been crashing at peak time weekends for about 3 weeks.

I have run 'top' and find lots of mysqld processes running even when it is a quiet time for visitors. My httpd error_log is reporting loads of 'Segmentation Fault [11]' errors.

As far as I can tell the crash is occurring simply because there are too many Apache or MySQL processes running and that is overloading the server.

I have tried tweaking my mysql settings using a my.cnf file but things haven't really got any better.

I'm looking for some help on what to do with this. I can provide MySQL variables and status results if anyone feels up to the challenge.

Briefly:
Apache version 1.3.19
PHP version 4.0.6
MySQL version 3.23.36
1Ghz CPU
512MB RAM
4GB SCSI HDD

Apache
max_connections = 150
max_per_child = 100

MySQL
max_connections = 80
key_buffer = 16M
join_buffer = 1M
sort buffer = 2M
table_cache = 1024
wait_timeout = 240
connect_timeout = 10
max_allowed_packet = 16M
max_connect_errors = 10
temp_table_size = 1M
thread stack = 655536
record buffer = 1M

Thanks

[Magena_Cadmar]

The table_cache = 1024 looks too high for that system.

In my_huge.cnf has table_cache = 512 and that is for settings for a system with 1-2Gig RAM running 4 processors.

Keep in mind that you need to have twice as many file descriptors available at the operating system level as what is indicated in the table cache.

If you do a show status, it should show you the opened tables and if that number is very large, then you should probably set the table cache to some high number. But the manual shows that as in comparison to 64.

This might not be your only issue, and it might not solve everything, but it certainly is something to look into. 😃