[Koha] Slow OPAC search

[Manoj Kummini]

Hello everyone,

I am new on this list, and have just begun, with a colleague, managing the
KOHA installation in our library that has about 10000 books and some
journals.

Our OPAC had become terribly slow and on exploring we noticed that a
mysqldump would result in a file of over 5GB. This large size, we realized,
was due to the cleanup_database.pl script not being called. After running
it for about a week with 

    --sessdays 7 --zebraqueue --list-invites --mail --merged --import --logs --searchhistory --restrictions

we have brought the mysqldump file to about 20-22 MB. 

While the speed of an OPAC search has improved considerably, it is still
slow. For example, as suggested by a reply to a similar question, I looked
at "Inspect Element > Network" and found that most queries of GET
opac-detail.pl and opac-search.pl take about 9-10 seconds.

Koha Version: 3.2202000
 (<meta name="generator" content="Koha 3.2202000" />)
running on a virtual machine (Debian 8.3). The base machine is Debian 9.0.
MySQL server 5.5.47-0+deb8u1 is on the VM itself.

Uses zebra.

On the virtual machine
# free -m 
gives 
             total       used       free     shared    buffers     cached
Mem:         16023       2042      13981        341        174       1290
-/+ buffers/cache:        576      15446
Swap:        16383          0      16383

We plan to upgrade to Koha 17.05 soon, but wanted to understand how to
speed up the search. Is moving to a stand-alone server the only solution?

I am new to this, so I might not have given you all the relevant
information about the problem. Please let me know if some other piece of
information will help.

Thank you.

With regards,
Manoj.

-- 
Manoj Kummini
Associate Professor, Mathematics, Chennai Mathematical Institute, 
Siruseri Tamilnadu 603103. India. 12 deg 50 min N, 80 deg 13 min E.
http://www.cmi.ac.in/~mkummini/index.html