Message-ID: <54F5BD29.6050705@gmail.com>
Date: Tue, 03 Mar 2015 16:54:49 +0300
From: Sergey Shchukin <shchukin.s.a@gmail.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64;
 rv:31.0) Gecko/20100101 Thunderbird/31.4.0
MIME-Version: 1.0
To: Jim Nasby <Jim.Nasby@BlueTreble.com>,
 Radovan Jablonovsky <radovan.jablonovsky@replicon.com>,
 pgsql-general@postgresql.org, pgsql-admin@postgresql.org
Subject: Re: [GENERAL] Re: [pgadmin-support] Issue with a hanging apply
 process
 on the replica db after vacuum works on primary
References: <54EC7FAE.8050003@gmail.com>
 <CAJYcdTtmk4bB9eXfEK_YSrqmXMmssJ3NVwTPXNw2UVrtCScoew@mail.gmail.com>
 <54EEBC5B.5000402@gmail.com> <54F03052.6050900@BlueTreble.com>
 <54F050D2.7030406@gmail.com> <54F0582F.7010402@gmail.com>
In-Reply-To: <54F0582F.7010402@gmail.com>
Content-Type: multipart/mixed; boundary="------------040608090303080804090900"
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

This is a multi-part message in MIME format.
--------------040608090303080804090900
Content-Type: multipart/alternative;
 boundary="------------000703000300080308020807"


--------------000703000300080308020807
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit

27.02.2015 14:42, Sergey Shchukin пишет:
> 27.02.2015 14:11, Sergey Shchukin пишет:
>> 27.02.2015 11:52, Jim Nasby пишет:
>>> On 2/26/15 12:25 AM, Sergey Shchukin wrote:
>>>> Hi Radovan !
>>>>
>>>> Thank you for the reply. The question is that this table is not a
>>>> subject for a massive updates/deletes.
>>>>
>>>> Is there any additional traces except from perf or pg_top to trace 
>>>> what
>>>> replica is doing at the particular moment when we are lagging in 
>>>> replay?
>>>> To see locks or spins or sleeps etc..
>>>
>>> Please don't top-post.
>>>
>>> What version is this? What is max_standby_streaming_delay set to?
>>>
>>>> Thank you!
>>>>
>>>> -
>>>>
>>>> Best regards,
>>>> Sergey Shchukin
>>>>
>>>> 24.02.2015 19:05, Radovan Jablonovsky пишет:
>>>>> This looks like more issue for pgsql-general mailing list.
>>>>>
>>>>> Possible solutions
>>>>> 1) Set specific autovacuum parameters on the big table. The 
>>>>> autovacuum
>>>>> could vacuum table on multiple runs based on the thresholds and cost
>>>>> settings
>>>>> Example of setting specific values of autovacuum and analyze for
>>>>> table. It should be adjusted for your system, work load, table 
>>>>> usage, etc:
>>>>> alter table "my_schema"."my_big_table" set (fillfactor = 80,
>>>>> autovacuum_enabled = true, autovacuum_vacuum_threshold = 200,
>>>>> autovacuum_analyze_threshold = 400, autovacuum_vacuum_scale_factor =
>>>>> 0.05, autovacuum_analyze_scale_factor = 0.005,
>>>>> autovacuum_vacuum_cost_delay = 10, autovacuum_vacuum_cost_limit = 
>>>>> 5000);
>>>>>
>>>>> 2) Could be to partition the large table on master site and vacuum it
>>>>> partition by partition.
>>>>>
>>>>> On Tue, Feb 24, 2015 at 6:42 AM, Sergey Shchukin
>>>>> <shchukin.s.a@gmail.com <mailto:shchukin.s.a@gmail.com>> wrote:
>>>>>
>>>>>     Hi all!
>>>>>
>>>>>     May someone help me with the issue in the apply process on the
>>>>>     replica. We have a stream replication and after vacuum stops
>>>>>     working with a big table we get a "freeze" in applying data on 
>>>>> the
>>>>>     replica database. It looks like this:
>>>>>
>>>>>     Tue Feb 24 15:04:51 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607136410456 SLAVE:79607136410456
>>>>>     Replay:79607136339456 :: REPLAY 69 KBytes (00:00:00.294485 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:52 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607137892672 SLAVE:79607137715392
>>>>>     Replay:79607137715392 :: REPLAY 173 KBytes (00:00:00.142605 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:53 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607139327776 SLAVE:79607139241816
>>>>>     Replay:79607139241816 :: REPLAY 84 KBytes (00:00:00.05223 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:54 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607141134776 SLAVE:79607141073344
>>>>>     Replay:79607141080032 :: REPLAY 54 KBytes (00:00:00.010603 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:55 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607143085176 SLAVE:79607143026440
>>>>>     Replay:79607143038040 :: REPLAY 46 KBytes (00:00:00.009506 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:56 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607145111280 SLAVE:79607145021384
>>>>>     Replay:79607145025664 :: REPLAY 83 KBytes (00:00:00.006795 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:57 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607146564424 SLAVE:79607146478336
>>>>>     Replay:79607146501264 :: REPLAY 61 KBytes (00:00:00.00701 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:58 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607148160680 SLAVE:79607148108352
>>>>>     Replay:79607147369320 :: REPLAY 773 KBytes (00:00:00.449702 
>>>>> seconds)
>>>>>     Tue Feb 24 15:04:59 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607150220688 SLAVE:79607150159632
>>>>>     Replay:79607150171312 :: REPLAY 48 KBytes (00:00:00.006594 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:00 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607152365360 SLAVE:79607152262696
>>>>>     Replay:79607152285240 :: REPLAY 78 KBytes (00:00:00.007042 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:02 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607154049848 SLAVE:79607154012624
>>>>>     Replay:79607153446800 :: REPLAY 589 KBytes (00:00:00.513637 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:03 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607155229992 SLAVE:79607155187864
>>>>>     Replay:79607155188312 :: REPLAY 41 KBytes (00:00:00.004773 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:04 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607156833968 SLAVE:79607156764128
>>>>>     Replay:79607156785488 :: REPLAY 47 KBytes (00:00:00.006846 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:05 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607158419848 SLAVE:79607158344856
>>>>>     Replay:79607158396352 :: REPLAY 23 KBytes (00:00:00.005228 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:06 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607160004776 SLAVE:79607159962400
>>>>>     Replay:79607159988888 :: REPLAY 16 KBytes (00:00:00.003162 
>>>>> seconds)
>>>>>     *--here apply process just stops*
>>>>>
>>>>>     Tue Feb 24 15:05:07 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607161592048 SLAVE:79607161550576
>>>>>     Replay:79607160986064 :: REPLAY 592 KBytes (00:00:00.398376 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:08 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607163272840 SLAVE:79607163231384
>>>>>     Replay:79607160986064 :: REPLAY 2233 KBytes (00:00:01.446759 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:09 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607164958632 SLAVE:79607164904448
>>>>>     Replay:79607160986064 :: REPLAY 3879 KBytes (00:00:02.497181 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:10 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607166819560 SLAVE:79607166777712
>>>>>     Replay:79607160986064 :: REPLAY 5697 KBytes (00:00:03.543107 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:11 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607168595280 SLAVE:79607168566536
>>>>>     Replay:79607160986064 :: REPLAY 7431 KBytes (00:00:04.589736 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:12 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607170372064 SLAVE:79607170252480
>>>>>     Replay:79607160986064 :: REPLAY 9166 KBytes (00:00:05.635918 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:13 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607171829480 SLAVE:79607171714144
>>>>>     Replay:79607160986064 :: REPLAY 10589 KBytes (00:00:06.688115 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:14 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607173152488 SLAVE:79607173152488
>>>>>     Replay:79607160986064 :: REPLAY 11881 KBytes (00:00:07.736993 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:15 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607174149968 SLAVE:79607174149968
>>>>>     Replay:79607160986064 :: REPLAY 12855 KBytes (00:00:08.78538 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:16 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607176448344 SLAVE:79607176252088
>>>>>     Replay:79607160986064 :: REPLAY 15100 KBytes (00:00:09.835184 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:17 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607177632216 SLAVE:79607177608224
>>>>>     Replay:79607160986064 :: REPLAY 16256 KBytes (00:00:10.926493 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:18 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607179432960 SLAVE:79607179378096
>>>>>     Replay:79607160986064 :: REPLAY 18015 KBytes (00:00:11.97989 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:19 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607180893384 SLAVE:79607180874256
>>>>>     Replay:79607160986064 :: REPLAY 19441 KBytes (00:00:13.028921 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:20 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607182596224 SLAVE:79607182552272
>>>>>     Replay:79607160986064 :: REPLAY 21104 KBytes (00:00:14.079497 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:21 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607183935312 SLAVE:79607183902592
>>>>>     Replay:79607160986064 :: REPLAY 22411 KBytes (00:00:15.127679 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:23 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607185165880 SLAVE:79607185094032
>>>>>     Replay:79607160986064 :: REPLAY 23613 KBytes (00:00:16.175132 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:24 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607187196920 SLAVE:79607187169368
>>>>>     Replay:79607160986064 :: REPLAY 25596 KBytes (00:00:17.221981 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:25 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607188943856 SLAVE:79607188885952
>>>>>     Replay:79607160986064 :: REPLAY 27302 KBytes (00:00:18.274362 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:26 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607190489400 SLAVE:79607190443160
>>>>>     Replay:79607160986064 :: REPLAY 28812 KBytes (00:00:19.319987 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:27 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607192089312 SLAVE:79607192054048
>>>>>     Replay:79607160986064 :: REPLAY 30374 KBytes (00:00:20.372305 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:28 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607193736800 SLAVE:79607193690056
>>>>>     Replay:79607160986064 :: REPLAY 31983 KBytes (00:00:21.421359 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:29 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607195968648 SLAVE:79607195901296
>>>>>     Replay:79607160986064 :: REPLAY 34163 KBytes (00:00:22.471334 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:30 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607197808840 SLAVE:79607197737720
>>>>>     Replay:79607160986064 :: REPLAY 35960 KBytes (00:00:23.52269 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:31 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607199571144 SLAVE:79607199495976
>>>>>     Replay:79607160986064 :: REPLAY 37681 KBytes (00:00:24.577615 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:32 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607201206104 SLAVE:79607201100392
>>>>>     Replay:79607160986064 :: REPLAY 39277 KBytes (00:00:25.624604 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:33 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607203174208 SLAVE:79607203111136
>>>>>     Replay:79607160986064 :: REPLAY 41199 KBytes (00:00:26.67059 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:34 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607204792888 SLAVE:79607204741600
>>>>>     Replay:79607160986064 :: REPLAY 42780 KBytes (00:00:27.719088 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:35 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607206453216 SLAVE:79607206409032
>>>>>     Replay:79607160986064 :: REPLAY 44401 KBytes (00:00:28.766647 
>>>>> seconds)
>>>>>     Tue Feb 24 15:05:36 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79607208225344 SLAVE:79607208142176
>>>>>     Replay:79607160986064 :: REPLAY 46132 KBytes (00:00:29.811434 
>>>>> seconds)
>>>>>
>>>>>
>>>>>     perf shows the following functions on the top
>>>>>     +  22.50%  postmaster  [kernel.kallsyms]  [k] 
>>>>> copy_user_generic_string
>>>>>     +   8.48%  postmaster  postgres           [.]
>>>>>     hash_search_with_hash_value
>>>>>
>>>>>
>>>>>     after 10 minutes or so the apply process continue to work
>>>>>
>>>>>     Tue Feb 24 15:13:25 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608758742560 SLAVE:79608758718008
>>>>>     Replay:79607160986064 :: REPLAY 1560309 KBytes (00:08:19.009653
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:26 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608759203608 SLAVE:79608759189680
>>>>>     Replay:79607160986064 :: REPLAY 1560759 KBytes (00:08:20.057877
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:27 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608759639680 SLAVE:79608759633224
>>>>>     Replay:79607160986064 :: REPLAY 1561185 KBytes (00:08:21.104723
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:28 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608760271200 SLAVE:79608760264128
>>>>>     Replay:79607160986064 :: REPLAY 1561802 KBytes (00:08:22.148546
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:30 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608760622920 SLAVE:79608760616656
>>>>>     Replay:79607160986064 :: REPLAY 1562145 KBytes (00:08:23.196645
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:31 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608761122040 SLAVE:79608761084584
>>>>>     Replay:79607160986064 :: REPLAY 1562633 KBytes (00:08:24.240653
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:32 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608761434200 SLAVE:79608761426080
>>>>>     Replay:79607160986064 :: REPLAY 1562938 KBytes (00:08:25.289429
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:33 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608761931008 SLAVE:79608761904808
>>>>>     Replay:79607160986064 :: REPLAY 1563423 KBytes (00:08:26.338498
>>>>>     seconds)
>>>>>     *--apply starts*
>>>>>     Tue Feb 24 15:13:34 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608762360568 SLAVE:79608762325712
>>>>>     Replay:79607163554680 :: REPLAY 1561334 KBytes (00:08:25.702423
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:35 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608762891224 SLAVE:79608762885928
>>>>>     Replay:79607166466488 :: REPLAY 1559008 KBytes (00:08:25.011046
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:36 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608763681920 SLAVE:79608763667256
>>>>>     Replay:79607167054056 :: REPLAY 1559207 KBytes (00:08:25.827531
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:37 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608764207088 SLAVE:79608764197744
>>>>>     Replay:79607175610296 :: REPLAY 1551364 KBytes (00:08:21.182428
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:38 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608764857920 SLAVE:79608764832432
>>>>>     Replay:79607183599632 :: REPLAY 1544197 KBytes (00:08:16.742467
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:39 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608765323360 SLAVE:79608765281408
>>>>>     Replay:79607186862176 :: REPLAY 1541466 KBytes (00:08:15.569874
>>>>>     seconds)
>>>>>     Tue Feb 24 15:13:40 MSK 2015 Stream:
>>>>>     MASTER-masterdb:79608765848240 SLAVE:79608765824520
>>>>>     Replay:79607186862176 :: REPLAY 1541978 KBytes (00:08:16.620932
>>>>>     seconds)
>>>>>
>>>>>
>>>>>     All this is a result of completion of  "vacuum verbose analyze
>>>>>     master_table" on the master site
>>>>>
>>>>>     Any help would be appreciated
>>>>>
>>>>>     --
>>>>>     Best regards,
>>>>>     Sergey Shchukin
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> -- 
>>>>>
>>>>> *Radovan Jablonovsky* | SaaS DBA | Phone 1-403-262-6519 (ext. 7256) |
>>>>> Fax 1-403-233-8046
>>>>>
>>>>>
>>>>>
>>>>
>>>
>>>
>> Hi Jim,
>>
>> The version is _PostgreSQL 9.3.6_ on x86_64 RHEL 6.6
>>
>> show max_standby_streaming_delay;
>>  max_standby_streaming_delay
>> -----------------------------
>>  30s
>>
>
> Again, after the vacuum finished on my table I got locks in apply 
> process on replica - *see lag_byte *
>
> masterdb01d/masterdb M # vacuum verbose rtable.rtable_uidl;
> INFO:  00000: vacuuming "rtable.rtable_uidl"
> LOCATION:  lazy_scan_heap, vacuumlazy.c:438
> INFO:  00000: scanned index "pk_rtable_uidl" to remove 6 row versions
> DETAIL:  CPU 240.80s/183.19u sec elapsed 703.85 sec.
> LOCATION:  lazy_vacuum_index, vacuumlazy.c:1335
> INFO:  00000: "rtable_uidl": removed 6 row versions in 6 pages
> DETAIL:  CPU 0.00s/0.00u sec elapsed 0.00 sec.
> LOCATION:  lazy_vacuum_heap, vacuumlazy.c:1169
> INFO:  00000: index "pk_rtable_uidl" now contains 3763411079 row 
> versions in 32755911 pages
> DETAIL:  0 index row versions were removed.
> 0 index pages have been deleted, 0 are currently reusable.
> CPU 0.00s/0.00u sec elapsed 0.00 sec.
> LOCATION:  lazy_cleanup_index, vacuumlazy.c:1387
> INFO:  00000: "rtable_uidl": found 6 removable, 1426488 nonremovable 
> row versions in 12734 out of 26047416 pages
> DETAIL:  0 dead row versions cannot be removed yet.
> There were 3 unused item pointers.
> 0 pages are entirely empty.
> CPU 241.37s/184.04u sec elapsed 705.45 sec.
> LOCATION:  lazy_scan_heap, vacuumlazy.c:1101
> VACUUM
> Time: 705685.954 ms
>
>
>
> masterdb01d/postgres M # \g
> -[ RECORD 1 ]----+------------------------------
> procpid          | 21487
> usesysid         | 16413
> usename          | repl
> application_name | walreceiver
> client_addr      |
> client_hostname  | masterdb01e
> client_port      | 35261
> backend_start    | 2015-02-27 13:02:27.203938+03
> state            | streaming
> sent_location    | 494B/CB30B530
> write_location   | 494B/CB30B530
> flush_location   | 494B/CB30B530
> replay_location  | *494B/A02B9070  <<< stopped here!1*
> sync_priority    | 0
> sync_state       | async
> total_lag_byte   | 721757376
> total_lag_nice   | 688 MB
> lag_byte         | 721757376
> lag_byte_nice    | 688 MB
>
> --
> Best regards,
> Sergey Shchukin

Hi All!

The issue is repeating still...

Some updates:

The size of the rtable_uidl is about 3 773 185 761 rows, quite a big

Before vacuum

  schema_size | schema_size_with_indexes
-------------+--------------------------
  199 GB      | 450 GB


rtabledb01e/rtabledb R # select * from pg_stat_all_tables where 
schemaname = 'rtable' and relname ='rtable_uidl';
-[ RECORD 1 ]-----+-----------
relid             | 16511
schemaname        | rtable
relname           | rtable_uidl
seq_scan          | 2
seq_tup_read      | 7546016953
idx_scan          | 3008619
idx_tup_fetch     | 44139478
n_tup_ins         | 0
n_tup_upd         | 0
n_tup_del         | 0
n_tup_hot_upd     | 0
n_live_tup        | 0
n_dead_tup        | 0
last_vacuum       | [null]
last_autovacuum   | [null]
last_analyze      | [null]
last_autoanalyze  | [null]
vacuum_count      | 0
autovacuum_count  | 0
analyze_count     | 0
autoanalyze_count | 0


Server spec: 2 Xeon E5-2660, 128GB RAM, disk subsystem

/dev/md2         /var/lib/pgsql/9.3/data
/dev/md3         /var/lib/pgsql/9.3/data/pg_xlog

raids are raid10 on top of 8x300GB intel SSD disks. Both raids use the 
same disks which are split into two parts (ex. md2:sdb1.., md3:sdb2...).

It should be mentioned that we see significant disc activity on replica 
DBS during the lockup of apply process (see disks_load.png)

OS: RHEL 6.6

Non default parameters - please check the attachment

After the vacuum
rtabledb01d/rtabledb M # vacuum verbose rtable.rtable_uidl;
INFO:  00000: vacuuming "rtable.rtable_uidl"
LOCATION:  lazy_scan_heap, vacuumlazy.c:438
INFO:  00000: scanned index "pk_rtable_uidl" to remove 3 row versions
DETAIL:  CPU 251.85s/165.91u sec elapsed 743.44 sec.
LOCATION:  lazy_vacuum_index, vacuumlazy.c:1335
INFO:  00000: "rtable_uidl": removed 3 row versions in 3 pages
DETAIL:  CPU 0.00s/0.00u sec elapsed 0.00 sec.
LOCATION:  lazy_vacuum_heap, vacuumlazy.c:1169
INFO:  00000: index "pk_rtable_uidl" now contains 3773254753 row 
versions in 32846328 pages
DETAIL:  0 index row versions were removed.
0 index pages have been deleted, 0 are currently reusable.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
LOCATION:  lazy_cleanup_index, vacuumlazy.c:1387
INFO:  00000: "rtable_uidl": found 3 removable, 2589545 nonremovable row 
versions in 20398 out of 26102225 pages
DETAIL:  0 dead row versions cannot be removed yet.
There were 0 unused item pointers.
0 pages are entirely empty.
CPU 253.20s/167.60u sec elapsed 749.44 sec.
LOCATION:  lazy_scan_heap, vacuumlazy.c:1101
VACUUM
Time: 750141.401 ms

     date
------------
  2015-03-03
(1 row)

        timetz
--------------------
  16:35:50.245227+03
(1 row)


Replay stopped on replicas
Tue Mar  3 16:35:12 MSK 2015 Stream: MASTER-rtabledb01d:81829147840528 
SLAVE:81829147765224 Replay:*81827156622184 *:: REPLAY 1944550 KBytes 
(00:07:49.816045 seconds)
Tue Mar  3 16:35:17 MSK 2015 Stream: MASTER-rtabledb01d:81829156031576 
SLAVE:81829155989056 Replay:*81827156622184 *:: REPLAY 1952549 KBytes 
(00:07:54.87067 seconds)
Tue Mar  3 16:35:22 MSK 2015 Stream: MASTER-rtabledb01d:81829165013208 
SLAVE:81829164949312 Replay:*81827156622184 *:: REPLAY 1961320 KBytes 
(00:07:59.927685 seconds)
Tue Mar  3 16:35:27 MSK 2015 Stream: MASTER-rtabledb01d:81829172941288 
SLAVE:81829172887000 Replay:*81827156622184 *:: REPLAY 1969062 KBytes 
(00:08:04.977663 seconds)
Tue Mar  3 16:35:32 MSK 2015 Stream: MASTER-rtabledb01d:81829181088512 
SLAVE:81829181063304 Replay:*81827156622184 *:: REPLAY 1977018 KBytes 
(00:08:10.033499 seconds)
Tue Mar  3 16:35:37 MSK 2015 Stream: MASTER-rtabledb01d:81829191442216 
SLAVE:81829191364192 Replay:*81827156622184 *:: REPLAY 1987129 KBytes 
(00:08:15.085862 seconds)
Tue Mar  3 16:35:42 MSK 2015 Stream: MASTER-rtabledb01d:81829204580736 
SLAVE:81829204444136 Replay:*81827156622184 *:: REPLAY 1999960 KBytes 
(00:08:20.13815 seconds)
Tue Mar  3 16:35:47 MSK 2015 Stream: MASTER-rtabledb01d:81829218243240 
SLAVE:81829218115272 Replay:*81827156622184 *:: REPLAY 2013302 KBytes 
(00:08:25.190515 seconds)

After the vacuum

schema_size | schema_size_with_indexes
-------------+--------------------------
  199 GB      | 450 GB

rtabledb01d/rtabledb M # select * from pg_stat_all_tables where 
schemaname = 'rtable' and relname ='rtable_uidl';
-[ RECORD 1 ]-----+------------------------------
relid             | 16511
schemaname        | rtable
relname           | rtable_uidl
seq_scan          | 1
seq_tup_read      | 14223712
idx_scan          | 1370133
idx_tup_fetch     | 28596708
n_tup_ins         | 12473820
n_tup_upd         | 0
n_tup_del         | 0
n_tup_hot_upd     | 0
n_live_tup        | 3771361529
n_dead_tup        | 0
last_vacuum       | 2015-03-03 16:27:23.231513+03
last_autovacuum   | 2015-03-03 07:11:50.891016+03
last_analyze      | 2015-02-26 19:50:28.097629+03
last_autoanalyze  | 2015-03-03 14:09:14.463953+03
vacuum_count      | 4
autovacuum_count  | 5
analyze_count     | 1
autoanalyze_count | 3

Thanks in advance!

Best regards,
Sergey Shchukin

--------------000703000300080308020807
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: 8bit

<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">27.02.2015 14:42, Sergey Shchukin
      пишет:<br>
    </div>
    <blockquote cite="mid:54F0582F.7010402@gmail.com" type="cite">
      <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
      <div class="moz-cite-prefix">27.02.2015 14:11, Sergey Shchukin
        пишет:<br>
      </div>
      <blockquote cite="mid:54F050D2.7030406@gmail.com" type="cite">
        <meta content="text/html; charset=utf-8"
          http-equiv="Content-Type">
        <div class="moz-cite-prefix">27.02.2015 11:52, Jim Nasby пишет:<br>
        </div>
        <blockquote cite="mid:54F03052.6050900@BlueTreble.com"
          type="cite">On 2/26/15 12:25 AM, Sergey Shchukin wrote: <br>
          <blockquote type="cite">Hi Radovan ! <br>
            <br>
            Thank you for the reply. The question is that this table is
            not a <br>
            subject for a massive updates/deletes. <br>
            <br>
            Is there any additional traces except from perf or pg_top to
            trace what <br>
            replica is doing at the particular moment when we are
            lagging in replay? <br>
            To see locks or spins or sleeps etc.. <br>
          </blockquote>
          <br>
          Please don't top-post. <br>
          <br>
          What version is this? What is max_standby_streaming_delay set
          to? <br>
          <br>
          <blockquote type="cite">Thank you! <br>
            <br>
            - <br>
            <br>
            Best regards, <br>
            Sergey Shchukin <br>
            <br>
            24.02.2015 19:05, Radovan Jablonovsky пишет: <br>
            <blockquote type="cite">This looks like more issue for
              pgsql-general mailing list. <br>
              <br>
              Possible solutions <br>
              1) Set specific autovacuum parameters on the big table.
              The autovacuum <br>
              could vacuum table on multiple runs based on the
              thresholds and cost <br>
              settings <br>
              Example of setting specific values of autovacuum and
              analyze for <br>
              table. It should be adjusted for your system, work load,
              table usage, etc: <br>
              alter table "my_schema"."my_big_table" set (fillfactor =
              80, <br>
              autovacuum_enabled = true, autovacuum_vacuum_threshold =
              200, <br>
              autovacuum_analyze_threshold = 400,
              autovacuum_vacuum_scale_factor = <br>
              0.05, autovacuum_analyze_scale_factor = 0.005, <br>
              autovacuum_vacuum_cost_delay = 10,
              autovacuum_vacuum_cost_limit = 5000); <br>
              <br>
              2) Could be to partition the large table on master site
              and vacuum it <br>
              partition by partition. <br>
              <br>
              On Tue, Feb 24, 2015 at 6:42 AM, Sergey Shchukin <br>
              &lt;<a moz-do-not-send="true"
                class="moz-txt-link-abbreviated"
                href="mailto:shchukin.s.a@gmail.com">shchukin.s.a@gmail.com</a>
              <a moz-do-not-send="true" class="moz-txt-link-rfc2396E"
                href="mailto:shchukin.s.a@gmail.com">&lt;mailto:shchukin.s.a@gmail.com&gt;</a>&gt;

              wrote: <br>
              <br>
                  Hi all! <br>
              <br>
                  May someone help me with the issue in the apply
              process on the <br>
                  replica. We have a stream replication and after vacuum
              stops <br>
                  working with a big table we get a "freeze" in applying
              data on the <br>
                  replica database. It looks like this: <br>
              <br>
                  Tue Feb 24 15:04:51 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607136410456 SLAVE:79607136410456 <br>
                  Replay:79607136339456 :: REPLAY 69 KBytes
              (00:00:00.294485 seconds) <br>
                  Tue Feb 24 15:04:52 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607137892672 SLAVE:79607137715392 <br>
                  Replay:79607137715392 :: REPLAY 173 KBytes
              (00:00:00.142605 seconds) <br>
                  Tue Feb 24 15:04:53 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607139327776 SLAVE:79607139241816 <br>
                  Replay:79607139241816 :: REPLAY 84 KBytes
              (00:00:00.05223 seconds) <br>
                  Tue Feb 24 15:04:54 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607141134776 SLAVE:79607141073344 <br>
                  Replay:79607141080032 :: REPLAY 54 KBytes
              (00:00:00.010603 seconds) <br>
                  Tue Feb 24 15:04:55 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607143085176 SLAVE:79607143026440 <br>
                  Replay:79607143038040 :: REPLAY 46 KBytes
              (00:00:00.009506 seconds) <br>
                  Tue Feb 24 15:04:56 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607145111280 SLAVE:79607145021384 <br>
                  Replay:79607145025664 :: REPLAY 83 KBytes
              (00:00:00.006795 seconds) <br>
                  Tue Feb 24 15:04:57 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607146564424 SLAVE:79607146478336 <br>
                  Replay:79607146501264 :: REPLAY 61 KBytes
              (00:00:00.00701 seconds) <br>
                  Tue Feb 24 15:04:58 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607148160680 SLAVE:79607148108352 <br>
                  Replay:79607147369320 :: REPLAY 773 KBytes
              (00:00:00.449702 seconds) <br>
                  Tue Feb 24 15:04:59 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607150220688 SLAVE:79607150159632 <br>
                  Replay:79607150171312 :: REPLAY 48 KBytes
              (00:00:00.006594 seconds) <br>
                  Tue Feb 24 15:05:00 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607152365360 SLAVE:79607152262696 <br>
                  Replay:79607152285240 :: REPLAY 78 KBytes
              (00:00:00.007042 seconds) <br>
                  Tue Feb 24 15:05:02 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607154049848 SLAVE:79607154012624 <br>
                  Replay:79607153446800 :: REPLAY 589 KBytes
              (00:00:00.513637 seconds) <br>
                  Tue Feb 24 15:05:03 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607155229992 SLAVE:79607155187864 <br>
                  Replay:79607155188312 :: REPLAY 41 KBytes
              (00:00:00.004773 seconds) <br>
                  Tue Feb 24 15:05:04 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607156833968 SLAVE:79607156764128 <br>
                  Replay:79607156785488 :: REPLAY 47 KBytes
              (00:00:00.006846 seconds) <br>
                  Tue Feb 24 15:05:05 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607158419848 SLAVE:79607158344856 <br>
                  Replay:79607158396352 :: REPLAY 23 KBytes
              (00:00:00.005228 seconds) <br>
                  Tue Feb 24 15:05:06 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607160004776 SLAVE:79607159962400 <br>
                  Replay:79607159988888 :: REPLAY 16 KBytes
              (00:00:00.003162 seconds) <br>
                  *--here apply process just stops* <br>
              <br>
                  Tue Feb 24 15:05:07 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607161592048 SLAVE:79607161550576 <br>
                  Replay:79607160986064 :: REPLAY 592 KBytes
              (00:00:00.398376 seconds) <br>
                  Tue Feb 24 15:05:08 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607163272840 SLAVE:79607163231384 <br>
                  Replay:79607160986064 :: REPLAY 2233 KBytes
              (00:00:01.446759 seconds) <br>
                  Tue Feb 24 15:05:09 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607164958632 SLAVE:79607164904448 <br>
                  Replay:79607160986064 :: REPLAY 3879 KBytes
              (00:00:02.497181 seconds) <br>
                  Tue Feb 24 15:05:10 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607166819560 SLAVE:79607166777712 <br>
                  Replay:79607160986064 :: REPLAY 5697 KBytes
              (00:00:03.543107 seconds) <br>
                  Tue Feb 24 15:05:11 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607168595280 SLAVE:79607168566536 <br>
                  Replay:79607160986064 :: REPLAY 7431 KBytes
              (00:00:04.589736 seconds) <br>
                  Tue Feb 24 15:05:12 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607170372064 SLAVE:79607170252480 <br>
                  Replay:79607160986064 :: REPLAY 9166 KBytes
              (00:00:05.635918 seconds) <br>
                  Tue Feb 24 15:05:13 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607171829480 SLAVE:79607171714144 <br>
                  Replay:79607160986064 :: REPLAY 10589 KBytes
              (00:00:06.688115 seconds) <br>
                  Tue Feb 24 15:05:14 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607173152488 SLAVE:79607173152488 <br>
                  Replay:79607160986064 :: REPLAY 11881 KBytes
              (00:00:07.736993 seconds) <br>
                  Tue Feb 24 15:05:15 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607174149968 SLAVE:79607174149968 <br>
                  Replay:79607160986064 :: REPLAY 12855 KBytes
              (00:00:08.78538 seconds) <br>
                  Tue Feb 24 15:05:16 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607176448344 SLAVE:79607176252088 <br>
                  Replay:79607160986064 :: REPLAY 15100 KBytes
              (00:00:09.835184 seconds) <br>
                  Tue Feb 24 15:05:17 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607177632216 SLAVE:79607177608224 <br>
                  Replay:79607160986064 :: REPLAY 16256 KBytes
              (00:00:10.926493 seconds) <br>
                  Tue Feb 24 15:05:18 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607179432960 SLAVE:79607179378096 <br>
                  Replay:79607160986064 :: REPLAY 18015 KBytes
              (00:00:11.97989 seconds) <br>
                  Tue Feb 24 15:05:19 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607180893384 SLAVE:79607180874256 <br>
                  Replay:79607160986064 :: REPLAY 19441 KBytes
              (00:00:13.028921 seconds) <br>
                  Tue Feb 24 15:05:20 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607182596224 SLAVE:79607182552272 <br>
                  Replay:79607160986064 :: REPLAY 21104 KBytes
              (00:00:14.079497 seconds) <br>
                  Tue Feb 24 15:05:21 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607183935312 SLAVE:79607183902592 <br>
                  Replay:79607160986064 :: REPLAY 22411 KBytes
              (00:00:15.127679 seconds) <br>
                  Tue Feb 24 15:05:23 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607185165880 SLAVE:79607185094032 <br>
                  Replay:79607160986064 :: REPLAY 23613 KBytes
              (00:00:16.175132 seconds) <br>
                  Tue Feb 24 15:05:24 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607187196920 SLAVE:79607187169368 <br>
                  Replay:79607160986064 :: REPLAY 25596 KBytes
              (00:00:17.221981 seconds) <br>
                  Tue Feb 24 15:05:25 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607188943856 SLAVE:79607188885952 <br>
                  Replay:79607160986064 :: REPLAY 27302 KBytes
              (00:00:18.274362 seconds) <br>
                  Tue Feb 24 15:05:26 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607190489400 SLAVE:79607190443160 <br>
                  Replay:79607160986064 :: REPLAY 28812 KBytes
              (00:00:19.319987 seconds) <br>
                  Tue Feb 24 15:05:27 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607192089312 SLAVE:79607192054048 <br>
                  Replay:79607160986064 :: REPLAY 30374 KBytes
              (00:00:20.372305 seconds) <br>
                  Tue Feb 24 15:05:28 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607193736800 SLAVE:79607193690056 <br>
                  Replay:79607160986064 :: REPLAY 31983 KBytes
              (00:00:21.421359 seconds) <br>
                  Tue Feb 24 15:05:29 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607195968648 SLAVE:79607195901296 <br>
                  Replay:79607160986064 :: REPLAY 34163 KBytes
              (00:00:22.471334 seconds) <br>
                  Tue Feb 24 15:05:30 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607197808840 SLAVE:79607197737720 <br>
                  Replay:79607160986064 :: REPLAY 35960 KBytes
              (00:00:23.52269 seconds) <br>
                  Tue Feb 24 15:05:31 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607199571144 SLAVE:79607199495976 <br>
                  Replay:79607160986064 :: REPLAY 37681 KBytes
              (00:00:24.577615 seconds) <br>
                  Tue Feb 24 15:05:32 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607201206104 SLAVE:79607201100392 <br>
                  Replay:79607160986064 :: REPLAY 39277 KBytes
              (00:00:25.624604 seconds) <br>
                  Tue Feb 24 15:05:33 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607203174208 SLAVE:79607203111136 <br>
                  Replay:79607160986064 :: REPLAY 41199 KBytes
              (00:00:26.67059 seconds) <br>
                  Tue Feb 24 15:05:34 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607204792888 SLAVE:79607204741600 <br>
                  Replay:79607160986064 :: REPLAY 42780 KBytes
              (00:00:27.719088 seconds) <br>
                  Tue Feb 24 15:05:35 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607206453216 SLAVE:79607206409032 <br>
                  Replay:79607160986064 :: REPLAY 44401 KBytes
              (00:00:28.766647 seconds) <br>
                  Tue Feb 24 15:05:36 MSK 2015 Stream: <br>
                  MASTER-masterdb:79607208225344 SLAVE:79607208142176 <br>
                  Replay:79607160986064 :: REPLAY 46132 KBytes
              (00:00:29.811434 seconds) <br>
              <br>
              <br>
                  perf shows the following functions on the top <br>
                  +  22.50%  postmaster  [kernel.kallsyms]  [k]
              copy_user_generic_string <br>
                  +   8.48%  postmaster  postgres           [.] <br>
                  hash_search_with_hash_value <br>
              <br>
              <br>
                  after 10 minutes or so the apply process continue to
              work <br>
              <br>
                  Tue Feb 24 15:13:25 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608758742560 SLAVE:79608758718008 <br>
                  Replay:79607160986064 :: REPLAY 1560309 KBytes
              (00:08:19.009653 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:26 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608759203608 SLAVE:79608759189680 <br>
                  Replay:79607160986064 :: REPLAY 1560759 KBytes
              (00:08:20.057877 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:27 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608759639680 SLAVE:79608759633224 <br>
                  Replay:79607160986064 :: REPLAY 1561185 KBytes
              (00:08:21.104723 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:28 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608760271200 SLAVE:79608760264128 <br>
                  Replay:79607160986064 :: REPLAY 1561802 KBytes
              (00:08:22.148546 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:30 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608760622920 SLAVE:79608760616656 <br>
                  Replay:79607160986064 :: REPLAY 1562145 KBytes
              (00:08:23.196645 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:31 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608761122040 SLAVE:79608761084584 <br>
                  Replay:79607160986064 :: REPLAY 1562633 KBytes
              (00:08:24.240653 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:32 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608761434200 SLAVE:79608761426080 <br>
                  Replay:79607160986064 :: REPLAY 1562938 KBytes
              (00:08:25.289429 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:33 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608761931008 SLAVE:79608761904808 <br>
                  Replay:79607160986064 :: REPLAY 1563423 KBytes
              (00:08:26.338498 <br>
                  seconds) <br>
                  *--apply starts* <br>
                  Tue Feb 24 15:13:34 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608762360568 SLAVE:79608762325712 <br>
                  Replay:79607163554680 :: REPLAY 1561334 KBytes
              (00:08:25.702423 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:35 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608762891224 SLAVE:79608762885928 <br>
                  Replay:79607166466488 :: REPLAY 1559008 KBytes
              (00:08:25.011046 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:36 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608763681920 SLAVE:79608763667256 <br>
                  Replay:79607167054056 :: REPLAY 1559207 KBytes
              (00:08:25.827531 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:37 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608764207088 SLAVE:79608764197744 <br>
                  Replay:79607175610296 :: REPLAY 1551364 KBytes
              (00:08:21.182428 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:38 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608764857920 SLAVE:79608764832432 <br>
                  Replay:79607183599632 :: REPLAY 1544197 KBytes
              (00:08:16.742467 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:39 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608765323360 SLAVE:79608765281408 <br>
                  Replay:79607186862176 :: REPLAY 1541466 KBytes
              (00:08:15.569874 <br>
                  seconds) <br>
                  Tue Feb 24 15:13:40 MSK 2015 Stream: <br>
                  MASTER-masterdb:79608765848240 SLAVE:79608765824520 <br>
                  Replay:79607186862176 :: REPLAY 1541978 KBytes
              (00:08:16.620932 <br>
                  seconds) <br>
              <br>
              <br>
                  All this is a result of completion of  "vacuum verbose
              analyze <br>
                  master_table" on the master site <br>
              <br>
                  Any help would be appreciated <br>
              <br>
                  -- <br>
                  Best regards, <br>
                  Sergey Shchukin <br>
              <br>
              <br>
              <br>
              <br>
              -- <br>
              <br>
              *Radovan Jablonovsky* | SaaS DBA | Phone 1-403-262-6519
              (ext. 7256) | <br>
              Fax 1-403-233-8046 <br>
              <br>
              <br>
              <br>
            </blockquote>
            <br>
          </blockquote>
          <br>
          <br>
        </blockquote>
        Hi Jim,<br>
        <br>
        The version is <u>PostgreSQL 9.3.6</u> on x86_64 RHEL 6.6 <br>
        <br>
        <small><font face="Courier New, Courier, monospace">show
            max_standby_streaming_delay;<br>
             max_standby_streaming_delay<br>
            -----------------------------<br>
             30s</font></small><br>
        <br>
      </blockquote>
      <br>
      <small><small><big><big><big><tt>Again, after the vacuum </tt></big></big></big></small></small><small><small><big><big><big><big><tt><small><small><small><small><big><big><big>finished

                                </big></big></big></small></small></small></small></tt></big><tt>on

                  my table I got locks in apply process on replica - <b>see

                    lag_byte </b></tt></big></big></big><font
            face="Courier New, Courier, monospace"><br>
            <br>
            masterdb01d/masterdb M # vacuum verbose rtable.rtable_uidl;<br>
            INFO:  00000: vacuuming "rtable.rtable_uidl"<br>
            LOCATION:  lazy_scan_heap, vacuumlazy.c:438<br>
            INFO:  00000: scanned index "pk_rtable_uidl" to remove 6 row
            versions<br>
            DETAIL:  CPU 240.80s/183.19u sec elapsed 703.85 sec.<br>
            LOCATION:  lazy_vacuum_index, vacuumlazy.c:1335<br>
            INFO:  00000: "rtable_uidl": removed 6 row versions in 6
            pages<br>
            DETAIL:  CPU 0.00s/0.00u sec elapsed 0.00 sec.<br>
            LOCATION:  lazy_vacuum_heap, vacuumlazy.c:1169<br>
            INFO:  00000: index "pk_rtable_uidl" now contains 3763411079
            row versions in 32755911 pages<br>
            DETAIL:  0 index row versions were removed.<br>
            0 index pages have been deleted, 0 are currently reusable.<br>
            CPU 0.00s/0.00u sec elapsed 0.00 sec.<br>
            LOCATION:  lazy_cleanup_index, vacuumlazy.c:1387<br>
            INFO:  00000: "rtable_uidl": found 6 removable, 1426488
            nonremovable row versions in 12734 out of 26047416 pages<br>
            DETAIL:  0 dead row versions cannot be removed yet.<br>
            There were 3 unused item pointers.<br>
            0 pages are entirely empty.<br>
            CPU 241.37s/184.04u sec elapsed 705.45 sec.<br>
            LOCATION:  lazy_scan_heap, vacuumlazy.c:1101<br>
            VACUUM<br>
            Time: 705685.954 ms<br>
            <br>
            <br>
            <br>
            masterdb01d/postgres M # \g<br>
            -[ RECORD 1 ]----+------------------------------<br>
            procpid          | 21487<br>
            usesysid         | 16413<br>
            usename          | repl<br>
            application_name | walreceiver<br>
            client_addr      | <br>
            client_hostname  | masterdb01e<br>
            client_port      | 35261<br>
            backend_start    | 2015-02-27 13:02:27.203938+03<br>
            state            | streaming<br>
            sent_location    | 494B/CB30B530<br>
            write_location   | 494B/CB30B530<br>
            flush_location   | 494B/CB30B530<br>
            replay_location  | <b>494B/A02B9070  &lt;&lt;&lt; stopped
              here!1</b><br>
            sync_priority    | 0<br>
            sync_state       | async<br>
            total_lag_byte   | 721757376<br>
            total_lag_nice   | 688 MB<br>
            lag_byte         | 721757376<br>
            lag_byte_nice    | 688 MB<br>
          </font></small></small><br>
      --<br>
      Best regards,<br>
      Sergey Shchukin<br>
    </blockquote>
    <br>
    Hi All!<br>
    <br>
    The issue is repeating still...<br>
    <br>
    Some updates:<br>
    <br>
    The size of the rtable_uidl is about 3 773 185 761 rows, quite a big<br>
    <br>
    Before vacuum<br>
    <br>
    <small><small><font face="Courier New, Courier, monospace"> schema_size

          | schema_size_with_indexes<br>
          -------------+--------------------------<br>
           199 GB      | 450 GB</font></small></small><br>
    <br>
    <br>
    <small><small><font face="Courier New, Courier, monospace">rtabledb01e/rtabledb

          R # select * from pg_stat_all_tables where schemaname =
          'rtable' and relname ='rtable_uidl';<br>
          -[ RECORD 1 ]-----+-----------<br>
          relid             | 16511<br>
          schemaname        | rtable<br>
          relname           | rtable_uidl<br>
          seq_scan          | 2<br>
          seq_tup_read      | 7546016953<br>
          idx_scan          | 3008619<br>
          idx_tup_fetch     | 44139478<br>
          n_tup_ins         | 0<br>
          n_tup_upd         | 0<br>
          n_tup_del         | 0<br>
          n_tup_hot_upd     | 0<br>
          n_live_tup        | 0<br>
          n_dead_tup        | 0<br>
          last_vacuum       | [null]<br>
          last_autovacuum   | [null]<br>
          last_analyze      | [null]<br>
          last_autoanalyze  | [null]<br>
          vacuum_count      | 0<br>
          autovacuum_count  | 0<br>
          analyze_count     | 0<br>
          autoanalyze_count | 0</font></small></small><br>
    <br>
    <br>
    <br>
    Server spec: 2 Xeon E5-2660, 128GB RAM, disk subsystem<br>
    <br>
    /dev/md2         /var/lib/pgsql/9.3/data<br>
    /dev/md3         /var/lib/pgsql/9.3/data/pg_xlog<br>
    <br>
    raids are raid10 on top of 8x300GB intel SSD disks. Both raids use
    the same disks which are split into two parts (ex. md2:sdb1..,
    md3:sdb2...).<br>
    <br>
    It should be mentioned that we see significant disc activity on
    replica DBS during the lockup of apply process (see disks_load.png)<br>
    <br>
    OS: RHEL 6.6<br>
    <br>
    Non default parameters - please check the attachment<br>
    <br>
    After the vacuum<br>
    <small><small><font face="Courier New, Courier, monospace">rtabledb01d/rtabledb

          M # vacuum verbose rtable.rtable_uidl;<br>
          INFO:  00000: vacuuming "rtable.rtable_uidl"<br>
          LOCATION:  lazy_scan_heap, vacuumlazy.c:438<br>
          INFO:  00000: scanned index "pk_rtable_uidl" to remove 3 row
          versions<br>
          DETAIL:  CPU 251.85s/165.91u sec elapsed 743.44 sec.<br>
          LOCATION:  lazy_vacuum_index, vacuumlazy.c:1335<br>
          INFO:  00000: "rtable_uidl": removed 3 row versions in 3 pages<br>
          DETAIL:  CPU 0.00s/0.00u sec elapsed 0.00 sec.<br>
          LOCATION:  lazy_vacuum_heap, vacuumlazy.c:1169<br>
          INFO:  00000: index "pk_rtable_uidl" now contains 3773254753
          row versions in 32846328 pages<br>
          DETAIL:  0 index row versions were removed.<br>
          0 index pages have been deleted, 0 are currently reusable.<br>
          CPU 0.00s/0.00u sec elapsed 0.00 sec.<br>
          LOCATION:  lazy_cleanup_index, vacuumlazy.c:1387<br>
          INFO:  00000: "rtable_uidl": found 3 removable, 2589545
          nonremovable row versions in 20398 out of 26102225 pages<br>
          DETAIL:  0 dead row versions cannot be removed yet.<br>
          There were 0 unused item pointers.<br>
          0 pages are entirely empty.<br>
          CPU 253.20s/167.60u sec elapsed 749.44 sec.<br>
          LOCATION:  lazy_scan_heap, vacuumlazy.c:1101<br>
          VACUUM<br>
          Time: 750141.401 ms</font></small></small><br>
    <br>
    <small><font face="Courier New, Courier, monospace">    date<br>
        ------------<br>
         2015-03-03<br>
        (1 row)<br>
        <br>
               timetz<br>
        --------------------<br>
         16:35:50.245227+03<br>
        (1 row)<br>
        <br>
      </font></small><br>
    Replay stopped on replicas<br>
    <small><font face="Courier New, Courier, monospace">Tue Mar  3
        16:35:12 MSK 2015 Stream: MASTER-rtabledb01d:81829147840528
        SLAVE:81829147765224 Replay:<b>81827156622184 </b>:: REPLAY
        1944550 KBytes (00:07:49.816045 seconds)<br>
        Tue Mar  3 16:35:17 MSK 2015 Stream:
        MASTER-rtabledb01d:81829156031576 SLAVE:81829155989056 Replay:<b>81827156622184
        </b>:: REPLAY 1952549 KBytes (00:07:54.87067 seconds)<br>
        Tue Mar  3 16:35:22 MSK 2015 Stream:
        MASTER-rtabledb01d:81829165013208 SLAVE:81829164949312 Replay:<b>81827156622184
        </b>:: REPLAY 1961320 KBytes (00:07:59.927685 seconds)<br>
        Tue Mar  3 16:35:27 MSK 2015 Stream:
        MASTER-rtabledb01d:81829172941288 SLAVE:81829172887000 Replay:<b>81827156622184
        </b>:: REPLAY 1969062 KBytes (00:08:04.977663 seconds)<br>
        Tue Mar  3 16:35:32 MSK 2015 Stream:
        MASTER-rtabledb01d:81829181088512 SLAVE:81829181063304 Replay:<b>81827156622184
        </b>:: REPLAY 1977018 KBytes (00:08:10.033499 seconds)<br>
        Tue Mar  3 16:35:37 MSK 2015 Stream:
        MASTER-rtabledb01d:81829191442216 SLAVE:81829191364192 Replay:<b>81827156622184
        </b>:: REPLAY 1987129 KBytes (00:08:15.085862 seconds)<br>
        Tue Mar  3 16:35:42 MSK 2015 Stream:
        MASTER-rtabledb01d:81829204580736 SLAVE:81829204444136 Replay:<b>81827156622184
        </b>:: REPLAY 1999960 KBytes (00:08:20.13815 seconds)<br>
        Tue Mar  3 16:35:47 MSK 2015 Stream:
        MASTER-rtabledb01d:81829218243240 SLAVE:81829218115272 Replay:<b>81827156622184
        </b>:: REPLAY 2013302 KBytes (00:08:25.190515 seconds)</font></small><br>
    <br>
    After the vacuum<br>
    <br>
     <small><small><font face="Courier New, Courier, monospace">schema_size

          | schema_size_with_indexes<br>
          -------------+--------------------------<br>
           199 GB      | 450 GB<br>
          <br>
          rtabledb01d/rtabledb M # select * from pg_stat_all_tables
          where schemaname = 'rtable' and relname ='rtable_uidl';<br>
          -[ RECORD 1 ]-----+------------------------------<br>
          relid             | 16511<br>
          schemaname        | rtable<br>
          relname           | rtable_uidl<br>
          seq_scan          | 1<br>
          seq_tup_read      | 14223712<br>
          idx_scan          | 1370133<br>
          idx_tup_fetch     | 28596708<br>
          n_tup_ins         | 12473820<br>
          n_tup_upd         | 0<br>
          n_tup_del         | 0<br>
          n_tup_hot_upd     | 0<br>
          n_live_tup        | 3771361529<br>
          n_dead_tup        | 0<br>
          last_vacuum       | 2015-03-03 16:27:23.231513+03<br>
          last_autovacuum   | 2015-03-03 07:11:50.891016+03<br>
          last_analyze      | 2015-02-26 19:50:28.097629+03<br>
          last_autoanalyze  | 2015-03-03 14:09:14.463953+03<br>
          vacuum_count      | 4<br>
          autovacuum_count  | 5<br>
          analyze_count     | 1<br>
          autoanalyze_count | 3<br>
        </font></small></small><br>
    Thanks in advance!<br>
    <br>
    Best regards,<br>
    Sergey Shchukin<br>
  </body>
</html>

--------------000703000300080308020807--

--------------040608090303080804090900
Content-Type: image/png;
 name="disks_load.png"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
 filename="disks_load.png"
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--------------040608090303080804090900
Content-Type: text/plain; charset=windows-1251;
 name="pg_non_default_params.txt"
Content-Transfer-Encoding: base64
Content-Disposition: attachment;
 filename="pg_non_default_params.txt"
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--------------040608090303080804090900
Content-Type: text/plain
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
MIME-Version: 1.0


-- 
Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-admin

--------------040608090303080804090900--