Re: Logical Replication 08P01 invalid memory alloc request size 1095736448

2025-07-06 Thread Tom Lane
Soni M  writes:
> We just do patch updates from 15.12 to 15.13, restart Logical Master and
> then Logical Replica, and suddenly Logical replication stops working.
> Streaming replication of both are fine.

This sounds like the same bug previously discussed in

https://www.postgresql.org/message-id/flat/680bdaf6-f7d1-4536-b580-05c2760c67c6%40deepbluecap.com

We'll have a fix in next month's quarterly releases.  For the moment,
you could either roll back to 15.12 or cherry-pick the fix at
commit fc0fb77c5.

regards, tom lane




Re: Logical Replication 08P01 invalid memory alloc request size 1095736448

2025-07-06 Thread Soni M
Thanks Tom, really appreciate it.

On Sun, Jul 6, 2025 at 10:12 PM Tom Lane  wrote:

> Soni M  writes:
> > We just do patch updates from 15.12 to 15.13, restart Logical Master and
> > then Logical Replica, and suddenly Logical replication stops working.
> > Streaming replication of both are fine.
>
> This sounds like the same bug previously discussed in
>
>
> https://www.postgresql.org/message-id/flat/680bdaf6-f7d1-4536-b580-05c2760c67c6%40deepbluecap.com
>
> We'll have a fix in next month's quarterly releases.  For the moment,
> you could either roll back to 15.12 or cherry-pick the fix at
> commit fc0fb77c5.
>
> regards, tom lane
>


-- 
Regards,

Soni Maula Harriz


Re: Logical Replication 08P01 invalid memory alloc request size 1095736448

2025-07-06 Thread Soni M
Some updates :
Checking at lsn_diff from 33D44/EF3219B0 to 33D45/373C2D98,
 pg_wal_lsn_diff
-
  1208620008
which is quite different to memory alloc request 1095736448.

Digging deeper into debug messages :

DEBUG,0,"find_in_dynamic_libpath: trying
""/usr/pgsql-15/lib/libpqwalreceiver""","","logical replication
worker",,0
DEBUG,0,"find_in_dynamic_libpath: trying
""/usr/pgsql-15/lib/libpqwalreceiver.so""","","logical replication
worker",,0
DEBUG,0,"InitPostgres","","logical replication worker",,0
DEBUG,0,"my backend ID is 3","","logical replication worker",,0
DEBUG,0,"StartTransaction(1) name: unnamed; blockState: DEFAULT; state:
INPROGRESS, xid/subid/cid: 0/1/0","","logical replication worker",,0
DEBUG,0,"CommitTransaction(1) name: unnamed; blockState: STARTED;
state: INPROGRESS, xid/subid/cid: 0/1/0","","logical replication
worker",,0
DEBUG,0,"StartTransaction(1) name: unnamed; blockState: DEFAULT; state:
INPROGRESS, xid/subid/cid: 0/1/0","","logical replication worker",,0
LOG,0,"logical replication apply worker for subscription
""consprod_subscription"" has started","","logical replication
worker",,0
DEBUG,0,"CommitTransaction(1) name: unnamed; blockState: STARTED;
state: INPROGRESS, xid/subid/cid: 0/1/0","","logical replication
worker",,0
DEBUG,0,"connecting to publisher using connection string
""host=10.2.5.43 port=5432 user=replicator
dbname=consprod""","","logical replication worker",,0
DEBUG,0,"StartTransaction(1) name: unnamed; blockState: DEFAULT; state:
INPROGRESS, xid/subid/cid: 0/1/0","","logical replication worker",,0
DEBUG,0,"CommitTransaction(1) name: unnamed; blockState: STARTED;
state: INPROGRESS, xid/subid/cid: 0/1/0","","logical replication
worker",,0
DEBUG,0,"logical replication apply worker for subscription
""consprod_subscription"" two_phase is DISABLED","","logical
replication worker",,0
DEBUG,0,"sending feedback (force 0) to recv 33D45/373C2D98, write
33D45/373C2D98, flush 33D45/373C2D98","","logical replication
worker",,0
DEBUG,0,"StartTransaction(1) name: unnamed; blockState: DEFAULT; state:
INPROGRESS, xid/subid/cid: 0/1/0","","logical replication worker",,0
DEBUG,0,"CommitTransaction(1) name: unnamed; blockState: STARTED;
state: INPROGRESS, xid/subid/cid: 0/1/0","","logical replication
worker",,0
DEBUG,0,"sending feedback (force 0) to recv 33D45/373C2D98, write
33D45/373C2D98, flush 33D45/373C2D98","","logical replication
worker",,0
ERROR,08P01,"could not receive data from WAL stream: ERROR:  invalid memory
alloc request size 1095736448","","logical replication worker",,0
DEBUG,0,"shmem_exit(1): 5 before_shmem_exit callbacks to
make","","logical replication worker",,0
DEBUG,0,"shmem_exit(1): 7 on_shmem_exit callbacks to
make","","logical replication worker",,0
DEBUG,0,"proc_exit(1): 1 callbacks to make","","logical
replication worker",,0
DEBUG,0,"exit(1)","","logical replication worker",,0
DEBUG,0,"shmem_exit(-1): 0 before_shmem_exit callbacks to
make","","logical replication worker",,0
DEBUG,0,"shmem_exit(-1): 0 on_shmem_exit callbacks to
make","","logical replication worker",,0
DEBUG,0,"proc_exit(-1): 0 callbacks to make","","logical
replication worker",,0
LOG,0,"background worker ""logical replication worker"" (PID 62160)
exited with exit code 1","","postmaster",,0

Please help

On Sun, Jul 6, 2025 at 4:40 PM Soni M  wrote:

> Hello All,
> we have this set up :
> Postgres 15.13 on RHEL 9.6
>
> Logical Master also have 2 streaming replica
> Logical Replica also have 1 streaming replica
>
> We just do patch updates from 15.12 to 15.13, restart Logical Master and
> then Logical Replica, and suddenly Logical replication stops working.
> Streaming replication of both are fine.
>
> The message on Logical Replica :
> ERROR,08P01,"could not receive data from WAL stream: ERROR:  invalid
> memory alloc request size 1095736448","","logical replication
> worker",,0
>
> The message on Logical Master :
> "START_REPLICATION", LOG,0,"starting logical decoding for slot
> ""_subscription""","Streaming transactions committing after 33D45/373C2D98,
> reading WAL from 33D44/EF3219B0.","START_REPLICATION SLOT
> ""_subscription"" LOGICAL 33D45/373C2D98 (proto_version '3',
> publication_names '""_publication""')",,,"_subscription","walsender",,0
>
> "START_REPLICATION", ,LOG,0,"logical decoding found consistent point
> at 33D44/EF3219B0","Logical decoding will begin using saved
> snapshot.","START_REPLICATION SLOT ""_subscription"" LOGICAL
> 33D45/373C2D98 (proto_version '3', publication_names
> '""_publication""')",,,"_subscription","walsender",,0
>
> Checking at error code 08P01 is protocol_violation.
>
> Is there any idea on what causing this and how t

Logical Replication 08P01 invalid memory alloc request size 1095736448

2025-07-06 Thread Soni M
Hello All,
we have this set up :
Postgres 15.13 on RHEL 9.6

Logical Master also have 2 streaming replica
Logical Replica also have 1 streaming replica

We just do patch updates from 15.12 to 15.13, restart Logical Master and
then Logical Replica, and suddenly Logical replication stops working.
Streaming replication of both are fine.

The message on Logical Replica :
ERROR,08P01,"could not receive data from WAL stream: ERROR:  invalid memory
alloc request size 1095736448","","logical replication worker",,0

The message on Logical Master :
"START_REPLICATION", LOG,0,"starting logical decoding for slot
""_subscription""","Streaming transactions committing after 33D45/373C2D98,
reading WAL from 33D44/EF3219B0.","START_REPLICATION SLOT
""_subscription"" LOGICAL 33D45/373C2D98 (proto_version '3',
publication_names '""_publication""')",,,"_subscription","walsender",,0

"START_REPLICATION", ,LOG,0,"logical decoding found consistent point at
33D44/EF3219B0","Logical decoding will begin using saved
snapshot.","START_REPLICATION SLOT ""_subscription"" LOGICAL
33D45/373C2D98 (proto_version '3', publication_names
'""_publication""')",,,"_subscription","walsender",,0

Checking at error code 08P01 is protocol_violation.

Is there any idea on what causing this and how to resolve ?

-- 
Regards,

Soni Maula Harriz