979273 - Postgres processed_crash table partitions must exist for all reports table partitions

Status: RESOLVED FIXED

(Socorro :: Database, task)

Description

[K Lars Lohn [:lars] [:klohn]]

we started saving processed crashes in PG on a certain date.  The partitioned table for the processed crashes doesn't exist prior to that date.  If a priority job comes in for a crash older than that date, the PG crash storage code will fail with a 'table not found' error.

The processor iterates through each crash storage system attempting to save the processed crash.  If the first crash storage system raises an exception, the loop is broken and the other crash storage systems don't even get tried.  This behavior is to prevent an inconsistent state between storage systems.  

Solution: create the PG 'processed_crash' table partitions for all the same partitions that the 'reports' table has.

Comment 1

[Selena Deckelmann :selenamarie :selena]

On prod:

breakpad=# \d processed_crashes
            Table "public.processed_crashes"
     Column      |           Type           | Modifiers
-----------------+--------------------------+-----------
 uuid            | uuid                     | not null
 processed_crash | json                     | not null
 date_processed  | timestamp with time zone |
Number of child tables: 116 (Use \d+ to list them.)

breakpad=# \d reports
                                        Table "public.reports"
       Column        |           Type           |                      Modifiers
---------------------+--------------------------+------------------------------------------------------
 id                  | integer                  | not null default nextval('reports_id_seq'::regclass)
 client_crash_date   | timestamp with time zone |
 date_processed      | timestamp with time zone |
 uuid                | character varying(50)    | not null
 product             | character varying(30)    |
 version             | character varying(16)    |
 build               | character varying(30)    |
 signature           | character varying(255)   |
 url                 | character varying(255)   |
 install_age         | integer                  |
 last_crash          | integer                  |
 uptime              | integer                  |
 cpu_name            | character varying(100)   |
 cpu_info            | character varying(100)   |
 reason              | character varying(255)   |
 address             | character varying(20)    |
 os_name             | character varying(100)   |
 os_version          | character varying(100)   |
 email               | character varying(100)   |
 user_id             | character varying(50)    |
 started_datetime    | timestamp with time zone |
 completed_datetime  | timestamp with time zone |
 success             | boolean                  |
 truncated           | boolean                  |
 processor_notes     | text                     |
 user_comments       | character varying(1024)  |
 app_notes           | character varying(1024)  |
 distributor         | character varying(20)    |
 distributor_version | character varying(20)    |
 topmost_filenames   | text                     |
 addons_checked      | boolean                  |
 flash_version       | text                     |
 hangid              | text                     |
 process_type        | text                     |
 release_channel     | text                     |
 productid           | text                     |
 exploitability      | text                     |
 flash_process_dump  | text                     |
 update_channel      | text                     |
Number of child tables: 115 (Use \d+ to list them.)

I ended up with one extra processed_crashes table earlier than the available reports tables, but that should be fine. 

Also ran same routine on stage.