Re: [GENERAL] Too long startup time after each crash.

22.12.2016, 04:21, "Adrian Klaver" <adrian.klaver@aklaver.com>:
> On 12/21/2016 02:17 PM, neos@olansoft.com wrote:
>>  Hi.
>>
>>  Perhaps i've sent it to wrong mail-list previously.
>>
>>  After the upgrade postgresql had a strange behavior. After the crash, server launch now take a long time, up to
30-40minutes sometimes. When connecting to the server (as well as in the log), the message - "the database system is
startingup". Startup process consumes 0.4 - 1% CPU. 
>
> What upgrade and how was it done?

By standard distro method (emerge in my case), and then pg_upgrade utility (according to pg docs). Upgrade was from 9.4
to9.5, and now 9.5.3 to 9.6.1. 

>
> What crash and what caused it?
>
> In particular why are you seeing multiple crashes?

Sorry, I did not specify a crash, i mean technical failure on the server - power failure or sometimes CPU overheating
("home"server problem). 

>
> What is the log just before the crash?

Nothing, as expected, since it's not postgresql crash.

>
> What is the log after the the "database system is starting up" part?

First - "the database system was interrupted; last known up at..."

then many times "database system is starting up" (one per connection attempt)

and after 5-40 min i get - database system was not properly shut down; automatic recovery in progress...redo starts
from...readyto accept connections, and, finally, everything is going well. 

(my logs in ru_RU locale, so do not place them here)

>
>>  Tried different version 9.5.[2,3,4], 9.6.1 (9.4 works fine). On the Internet this problem is not met, perhaps the
problemlies in a combination of different factors (configuration database, Linux kernel), may be my config is wrong? 
>
> Do you mean you searched for this problem description and could not find it?

Yes, it is, my fault, Google translator is not always saving me from mistakes.

>
>>  Technical details:
>>
>>  Gentoo Linux 4.4.2-hardened
>>  current pg version 9.6.1
>>
>>  ## pg config:
>>
>>  listen_addresses = '*'
>>  port = 5432
>>  max_connections = 500
>>  shared_buffers = 2048MB
>>  temp_buffers = 16MB
>>  work_mem = 64MB
>>  maintenance_work_mem = 512MB
>>  max_stack_depth = 4MB
>>  dynamic_shared_memory_type = posix
>>  wal_level = replica
>>  wal_buffers = 8MB
>>  checkpoint_completion_target = 0.85
>>  archive_mode = on
>>  archive_command = 'gzip < %p > /home/pgbasebackup/xlogs/%f.gz'
>>  max_wal_senders = 3
>>  effective_cache_size = 9GB
>>  log_min_duration_statement = 1500
>>  log_timezone = 'Asia/Yekaterinburg'
>>  datestyle = 'iso, dmy'
>>  timezone = 'Asia/Yekaterinburg'
>>  lc_messages = 'ru_RU.UTF-8'
>>  lc_monetary = 'ru_RU.UTF-8'
>>  lc_numeric = 'ru_RU.UTF-8'
>>  lc_time = 'ru_RU.UTF-8'
>>  default_text_search_config = 'pg_catalog.russian'
>>  plperl.on_init = 'use utf8; use re; package utf8; require "utf8_heavy.pl";'
>>
>>  ## end of config
>>
>>  Sorry for my eng :)
>
> --
> Adrian Klaver
> adrian.klaver@aklaver.com