Postgres disk activity

[spurrellian]

Moderator Edit: This thread has been split from another - https://support.nagios.com/forum/viewto ... 16&t=42852
In the future, please create a new thread and link to the old one instead of adding on.

I have similar issue to Fred so will post as well?

I have 1700 hosts and 15000 services

Code: Select all

           objectname            | objecttype | reltuples |  size
---------------------------------+------------+-----------+---------
 xi_meta                         | r          |      1253 | 4680 kB
 xi_incidents                    | r          |     14446 | 2816 kB
 xi_meta_pkey                    | i          |      1253 | 600 kB
 xi_meta_keyname_idx             | i          |      1253 | 544 kB
 pg_proc                         | r          |      2220 | 432 kB
 pg_toast_197804                 | t          |       164 | 432 kB
 xi_auditlog_source              | i          |      1830 | 384 kB
 xi_usermeta                     | r          |      2543 | 368 kB
 pg_attribute                    | r          |      2397 | 368 kB
 event_time                      | i          |       801 | 360 kB
 xi_auditlog_user                | i          |      1830 | 352 kB
 xi_meta_metatype_id_idx         | i          |      1253 | 352 kB
 xi_auditlog_ip_address          | i          |      1830 | 344 kB
 pg_depend                       | r          |      5480 | 328 kB
 xi_auditlog                     | r          |      1830 | 296 kB
 xi_meta_metaobj_id_idx          | i          |      1253 | 280 kB
 xi_auditlog_type                | i          |      1830 | 280 kB
 pg_depend_reference_index       | i          |      5480 | 272 kB
 xi_events_pkey                  | i          |       801 | 264 kB
 pg_depend_depender_index        | i          |      5480 | 256 kB
 pg_statistic                    | r          |       425 | 240 kB
 pg_proc_proname_args_nsp_index  | i          |      2220 | 208 kB
 xi_auditlog_log_time            | i          |      1830 | 200 kB
 xi_auditlog_pkey                | i          |      1830 | 200 kB
 pg_toast_2618                   | t          |       115 | 200 kB
 pg_toast_2619                   | t          |        35 | 184 kB
 xi_events                       | r          |       801 | 176 kB
 pg_description                  | r          |      2403 | 168 kB
 pg_attribute_relid_attnam_index | i          |      2397 | 160 kB
 xi_usermeta_user_id_key         | i          |      2543 | 152 kB
 pg_description_o_c_o_index      | i          |      2403 | 128 kB
 pg_toast_197833                 | t          |        34 | 128 kB
 pg_attribute_relid_attnum_index | i          |      2397 | 112 kB
 pg_operator                     | r          |       704 | 104 kB
 xi_usermeta_autoload_idx        | i          |      2543 | 80 kB
 xi_usermeta_pkey                | i          |      2543 | 80 kB
 pg_rewrite                      | r          |        86 | 72 kB
 pg_class                        | r          |       305 | 64 kB
 pg_proc_oid_index               | i          |      2220 | 64 kB
(39 rows)

[mcapra]

Not sure if you've been watching the other thread, but this is probably due to the changes mentioned by @tgriep:

In the newer versions if Nagios 5.x.x, there were some changes done for how the Global Event handlers are run on the system.
The new way it runs is to add a new event to a table called xi_events which in your case it in the Postgres database.
Then a cron job runs, looks at the xi_events table, gets the event and runs it.
That process also uses the xi_meta and the xi_eventqueue tables for storing temporary data, etc. so that is why you are seeing more disk writes for the Postgres database.