3

Ok so i have a rails application on our server and I have thinking shinx installed. Every day at 12:30am I run a cron job that truncates the account and contact tables and reinserts them...here is my crons

30 0 * * * /bin/bash -l -c 'cd /var/www/active && script/rails runner -e production '\''Account.db_insert'\$
34 0 * * * /bin/bash -l -c 'cd /var/www/active && RAILS_ENV=production bundle exec rake ts:index --silent'
34 0 * * * /bin/bash -l -c 'cd /var/www/active && RAILS_ENV=production bundle exec rake ts:restart --silent'

As you can see the crons executes the Account.db_insert which reinserts contacts and accounts but everyday I get an email with an error....here are the two that i got recently....

Yesterdays Error

using config file '/var/www/active/config/production.sphinx.conf'...
indexing index 'contact_core'...
FATAL: failed to lock /var/www/active/db/sphinx/production/contact_core.spl: Resource temporarily unavailable, will not index. Try --rotate option.

Todays Error

using config file '/var/www/active/config/production.sphinx.conf'...
indexing index 'contact_core'...
collected 403 docs, 0.0 MB
sorted 0.1 Mhits, 100.0% done
ERROR: index 'contact_core': rename /var/www/active/db/sphinx/production/contact_core.tmp.spl to /var/www/active/db/sphinx/production/contact_core.new.spl failed: No such file or directory.
total 403 docs, 25092 bytes
total 0.069 sec, 361571 bytes/sec, 5807.16 docs/sec
distributed index 'contact' can not be directly indexed; skipping.
total 5 reads, 0.001 sec, 133.7 kb/call avg, 0.2 msec/call avg
total 10 writes, 0.012 sec, 177.8 kb/call avg, 1.2 msec/call avg

Any Idea what I am doing wrong....

UPDATE

After the suggestion to run 

 /usr/local/bin/indexer --config /etc/sphinxsearch/sphinx.conf --all --rotate

I got this error 

 using config file '/etc/sphinxsearch/sphinx.conf'...
 ERROR: unknown key name 'client_timeout' in /etc/sphinxsearch/sphinx.conf line 591 col 16.
 FATAL: failed to parse config file '/etc/sphinxsearch/sphinx.conf'.
Matt Elhotiby
  • 43,028
  • 85
  • 218
  • 321
  • 1
    You're using incorrect conf path. Use yours, not blindly copy what people suggest. `/usr/local/bin/indexer --config /var/www/active/config/production.sphinx.conf --all --rotate` – phil pirozhkov May 16 '13 at 14:47

1 Answers1

4

As the error message indicates, you might want to run the indexer with the --rotate switch.

I, too, use Thinking Sphinx but found it easier to just use the Sphinx indexer directly. I have Thinking Sphinx generate my config for me, then I can just call out to searchd and indexer directly, in my cron:

*/30 * * * * /usr/local/bin/indexer --config /etc/sphinxsearch/sphinx.conf --all --rotate

I have foundthis to be more helpful than going through the Rake tasks. In addition, its much faster as the whole Rails stack doesnt have to be loaded.

Cody Caughlan
  • 32,456
  • 5
  • 63
  • 68
  • Does this replace the two 34 0 * * * /bin/bash -l -c 'cd /var/www/active && RAILS_ENV=production bundle exec rake ts:index --silent' 34 0 * * * /bin/bash -l -c 'cd /var/www/active && RAILS_ENV=production bundle exec rake ts:restart --silent' – Matt Elhotiby Oct 27 '11 at 13:43
  • so it would be like */34 * * * * /usr/local/bin/indexer --config /etc/sphinxsearch/sphinx.conf --all --rotate since the two thinking sphinx rake taskes run at 12:34 – Matt Elhotiby Oct 27 '11 at 13:45
  • Your Sphinx config is in a different place than mine, it looks like yours it at: `/var/www/active/config/production.sphinx.conf`. So update the `--config` in your command(s) accordingly. – Cody Caughlan Oct 27 '11 at 20:51