By justin3 on

Using page cache mode = standard, i'm noticing our cache_page table is consistently filling up with tons of overhead.

Using phpMyAdmin
cache_page table size = 122mb overhead = 117mb

after Optimize
cache_page table size = 5mb overhead = 0mb

but then it will quickly fill up to previous levels... is this common? should we just have a cron script optimize the table nightly? or is this a problematic bug?

thanks

⋅ Categories: Drupal 6.x

Comments

dman’s picture

dman commented

Well, I learnt what MySQL overhead is today. I guess it follows that the cache table is full of it.
I haven't heard anyone ever say that running a cron OPTIMIZE daily is a bad thing. But I have often wondered why it's not a built-in feature of MySQL in the first place and we have to roll little external scripts to do so...

.dan. is the New Zealand Drupal Developer working on Government Web Standards

gpk’s picture

gpk commented

>our cache_page table is consistently filling up with tons of overhead
Probably what is happening is that it is filling up, then it is emptied (e.g. when a node is updated), and until such time as it fills up again the previously used space becomes "overhead".

gpk
----
www.alexoria.co.uk

justin3’s picture

justin3 commented

so if i read you right, running a nightly cron to optimize isn't really necessary. it'll may clean a little disk space, but won't have much of an effect on speed of page loads. esta cierto?

thanks

gpk’s picture

gpk commented

>isn't really necessary
Probably not, but it's probably not doing any harm. You'd have to run some performance tests to discover the impact. However running scheduled optimize on the entirety of a large database might not be such a good idea, since I think each table gets locked before it is optimized, and if a table takes a while to sort out then the whole app could hang. See e.g. http://forums.mysql.com/read.php?21,132924,132924, http://www.google.co.uk/search?hl=en&q=mysql+optimize+lock&btnG=Google+S...

gpk
----
www.alexoria.co.uk

preper’s picture

preper commented

I have the same problem. After truncate the cache_page table I optimize the table but overload problem is still continue.

I am using 6.24 core and 6.18 boost for caching.

What can be the solution? What did you do for that issue?