log_min_messages per backend type
Hi,
Sometimes you need to inspect some debug messages from autovacuum worker but
you cannot apply the same setting for backends (that could rapidly fill the log
file). This proposal aims to change log_min_messages to have different log
levels depending on which backend type the message comes from.
The syntax was changed from enum to string and it accepts a list of elements.
Instead of enum, it now accepts a comma-separated list of elements (string).
Each element is LOGLEVEL:BACKENDTYPE. It still accepts the old syntax (single
log level for backward compatibility. In this case, it sets all backend types
to the informed log level. For the list, the default log level (WARNING) is
used for the backend types that are not specified.
SET log_min_messages TO 'debug2:checkpointer, debug1:autovacuum';
SHOW log_min_messages;
log_min_messages
----------------------------------------
debug2:checkpointer, debug1:autovacuum
(1 row)
In the above example, it sets log level DEBUG2 to checkpointer process and
DEBUG1 to autovacuum launcher and autovacuum worker processes. The other
processes keep WARNING (default) as log level.
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On Wed, 18 Dec 2024 at 00:15, Euler Taveira <euler@eulerto.com> wrote:
Hi,
Sometimes you need to inspect some debug messages from autovacuum worker but
you cannot apply the same setting for backends (that could rapidly fill the log
file). This proposal aims to change log_min_messages to have different log
levels depending on which backend type the message comes from.
Hi! Looks like a sane proposal. Correct me if I'm wrong, but with your
patch it is not possible to configure something like "everybody ERROR,
but autovac DEBUG5"? I mean,
set log_min_messages to 'ERROR:default, DEBUG5:autovacuum' (not
specifying all possible cases?)
On Thu, Dec 19, 2024, at 1:13 AM, Kirill Reshke wrote:
Hi! Looks like a sane proposal. Correct me if I'm wrong, but with your
patch it is not possible to configure something like "everybody ERROR,
but autovac DEBUG5"? I mean,
set log_min_messages to 'ERROR:default, DEBUG5:autovacuum' (not
specifying all possible cases?)
No. Good point. I forgot to mention that ALL as backend type could be added.
--
Euler Taveira
EDB https://www.enterprisedb.com/
Hello Euler,
On 2024-Dec-17, Euler Taveira wrote:
Sometimes you need to inspect some debug messages from autovacuum worker but
you cannot apply the same setting for backends (that could rapidly fill the log
file). This proposal aims to change log_min_messages to have different log
levels depending on which backend type the message comes from.The syntax was changed from enum to string and it accepts a list of elements.
Instead of enum, it now accepts a comma-separated list of elements (string).
Each element is LOGLEVEL:BACKENDTYPE.
This format seems unintuitive. I would have thought you do it the other
way around, "backendtype:loglevel" ... that seems more natural because
it's like assigning the 'loglevel' value to the 'backendtype' element.
SET log_min_messages TO 'checkpointer:debug2, autovacuum:debug1';
I dislike the array of names in variable.c. We already have an array in
launch_backend.c (child_process_kinds), plus GetBackendTypeDesc in
miscinit.c. Maybe not for this patch to clean up though.
I think it should be acceptable to configure one global setting with
exceptions for particular backend types:
log_min_messages = WARNING, autovacuum:DEBUG1
Right now I think the code only accepts the unadorned log level if there
are no other items in the list. I think the proposal downthread is to
use the keyword ALL for this,
log_min_messages = all:WARNING, autovacuum:DEBUG1 # I don't like this
but I think it's inferior, because then "all" is not really "all", and I
think it would be different if I had said
log_min_messages = autovacuum:DEBUG1, all:WARNING # I don't like this
because it looks like the "all" entry should override the one I set for
autovacuum before, which frankly would not make sense to me.
So I think these two lines,
log_min_messages = WARNING, autovacuum:DEBUG1
log_min_messages = autovacuum:DEBUG1, WARNING
should behave identically and mean "set the level for autovacuum to
DEBUG1, and to any other backend type to WARNING.
Also, I think it'd be better to reject duplicates in the list. Right
now it looks like the last entry for one backend type overrides prior
ones. I mean
log_min_messages = autovacuum:DEBUG1, autovacuum:ERROR
would set autovacuum to error, but that might be mistake prone if your
string is long. So implementation-wise I suggest to initialize the
whole newlogminmsgs array to -1, then scan the list of entries (saving
an entry without backend type as the one to use later and) setting every
backend type to the number specified; if we see trying to set a value
that's already different from -1, throw error. After scanning the whole
log_min_messages array, we scan the newlogminmsgs and set any entries
that are still -1 to the value that we saved before.
The new code in variable.c should be before the /* DATESTYLE */ comment
rather than at the end of the file.
You still have many XXX comments. Also, postgresql.conf should list the
valid values for backendtype, as well as show an example of a valid
setting. Please don't use ALLCAPS backend types in the docs, this looks
ugly:
+ Valid <literal>BACKENDTYPE</literal> values are <literal>ARCHIVER</literal>, + <literal>AUTOVACUUM</literal>, <literal>BACKEND</literal>, + <literal>BGWORKER</literal>, <literal>BGWRITER</literal>, + <literal>CHECKPOINTER</literal>, <literal>LOGGER</literal>, + <literal>SLOTSYNCWORKER</literal>, <literal>WALRECEIVER</literal>, + <literal>WALSENDER</literal>, <literal>WALSUMMARIZER</literal>, and + <literal>WALWRITER</literal>.
--
Ãlvaro Herrera Breisgau, Deutschland â https://www.EnterpriseDB.com/
"No necesitamos banderas
No reconocemos fronteras" (Jorge González)
On Wed, Feb 5, 2025, at 3:51 PM, Álvaro Herrera wrote:
On 2024-Dec-17, Euler Taveira wrote:
Sometimes you need to inspect some debug messages from autovacuum worker but
you cannot apply the same setting for backends (that could rapidly fill the log
file). This proposal aims to change log_min_messages to have different log
levels depending on which backend type the message comes from.The syntax was changed from enum to string and it accepts a list of elements.
Instead of enum, it now accepts a comma-separated list of elements (string).
Each element is LOGLEVEL:BACKENDTYPE.This format seems unintuitive. I would have thought you do it the other
way around, "backendtype:loglevel" ... that seems more natural because
it's like assigning the 'loglevel' value to the 'backendtype' element.
SET log_min_messages TO 'checkpointer:debug2, autovacuum:debug1';
Alvaro, thanks for your feedback. Your suggestion makes sense to me.
I dislike the array of names in variable.c. We already have an array in
launch_backend.c (child_process_kinds), plus GetBackendTypeDesc in
miscinit.c. Maybe not for this patch to clean up though.
I thought about using child_process_kinds but two details made me give
up on the idea: (a) multiple names per backend type (for example,
B_BACKEND, B_DEAD_END_BACKEND, B_STANDALONE_BACKEND) and (b) spaces into
names. Maybe we should have a group into child_process_kinds but as you
said it seems material for another patch.
I think it should be acceptable to configure one global setting with
exceptions for particular backend types:log_min_messages = WARNING, autovacuum:DEBUG1
Right now I think the code only accepts the unadorned log level if there
are no other items in the list. I think the proposal downthread is to
use the keyword ALL for this,log_min_messages = all:WARNING, autovacuum:DEBUG1 # I don't like this
but I think it's inferior, because then "all" is not really "all", and I
think it would be different if I had saidlog_min_messages = autovacuum:DEBUG1, all:WARNING # I don't like this
because it looks like the "all" entry should override the one I set for
autovacuum before, which frankly would not make sense to me.
Good point. After reflection, I agree that "all" is not a good keyword.
This patch turns backend type as optional so WARNING means apply this
log level as a final step to the backend types that are not specified in
the list.
So I think these two lines,
log_min_messages = WARNING, autovacuum:DEBUG1
log_min_messages = autovacuum:DEBUG1, WARNINGshould behave identically and mean "set the level for autovacuum to
DEBUG1, and to any other backend type to WARNING.
Done.
Also, I think it'd be better to reject duplicates in the list. Right
now it looks like the last entry for one backend type overrides prior
ones. I meanlog_min_messages = autovacuum:DEBUG1, autovacuum:ERROR
would set autovacuum to error, but that might be mistake prone if your
string is long. So implementation-wise I suggest to initialize the
whole newlogminmsgs array to -1, then scan the list of entries (saving
an entry without backend type as the one to use later and) setting every
backend type to the number specified; if we see trying to set a value
that's already different from -1, throw error. After scanning the whole
log_min_messages array, we scan the newlogminmsgs and set any entries
that are still -1 to the value that we saved before.The new code in variable.c should be before the /* DATESTYLE */ comment
rather than at the end of the file.
It was added into the MISCELLANEOUS section.
You still have many XXX comments. Also, postgresql.conf should list the
valid values for backendtype, as well as show an example of a valid
setting. Please don't use ALLCAPS backend types in the docs, this looks
ugly:+ Valid <literal>BACKENDTYPE</literal> values are <literal>ARCHIVER</literal>, + <literal>AUTOVACUUM</literal>, <literal>BACKEND</literal>, + <literal>BGWORKER</literal>, <literal>BGWRITER</literal>, + <literal>CHECKPOINTER</literal>, <literal>LOGGER</literal>, + <literal>SLOTSYNCWORKER</literal>, <literal>WALRECEIVER</literal>, + <literal>WALSENDER</literal>, <literal>WALSUMMARIZER</literal>, and + <literal>WALWRITER</literal>.
Done.
Just to recap what was changed:
- patch was rebased
- backend type is optional and means all unspecified backend types
- generic log level (without backend type) is mandatory and the order it
ppears is not important (it is applied for the remaining backend types)
- fix Windows build
- new tests to cover the changes
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On 2025-03-04 Tu 7:33 PM, Euler Taveira wrote:
I think it should be acceptable to configure one global setting with
exceptions for particular backend types:log_min_messages = WARNING, autovacuum:DEBUG1
Right now I think the code only accepts the unadorned log level if there
are no other items in the list. I think the proposal downthread is to
use the keyword ALL for this, log_min_messages = all:WARNING, autovacuum:DEBUG1  # I don't like this
but I think it's inferior, because then "all" is not really "all", and I
think it would be different if I had said log_min_messages = autovacuum:DEBUG1, all:WARNING  # I don't like this
because it looks like the "all" entry should override the one I set for
autovacuum before, which frankly would not make sense to me.Good point. After reflection, I agree that "all" is not a good keyword.
This patch turns backend type as optional so WARNING means apply this
log level as a final step to the backend types that are not specified in
the list.So I think these two lines,
log_min_messages = WARNING, autovacuum:DEBUG1
log_min_messages = autovacuum:DEBUG1, WARNINGshould behave identically and mean "set the level for autovacuum to
DEBUG1, and to any other backend type to WARNING.Done.
Just bikeshedding a bit ...
I'm not mad keen on this design. I think the value should be either a
single setting like "WARNING" or a set of type:setting pairs. I agree
that "all" is a bad name, but I think "default" would make sense.
I can live with it but I think this just looks a bit odd.
cheers
andrew
--
Andrew Dunstan
EDB:https://www.enterprisedb.com
On 2025/03/05 9:33, Euler Taveira wrote:
+Â Â Â Â Â Â Â Valid <literal>BACKENDTYPE</literal> values are <literal>ARCHIVER</literal>, +Â Â Â Â Â Â Â <literal>AUTOVACUUM</literal>, <literal>BACKEND</literal>, +Â Â Â Â Â Â Â <literal>BGWORKER</literal>, <literal>BGWRITER</literal>, +Â Â Â Â Â Â Â <literal>CHECKPOINTER</literal>, <literal>LOGGER</literal>, +Â Â Â Â Â Â Â <literal>SLOTSYNCWORKER</literal>, <literal>WALRECEIVER</literal>, +Â Â Â Â Â Â Â <literal>WALSENDER</literal>, <literal>WALSUMMARIZER</literal>, and +Â Â Â Â Â Â Â <literal>WALWRITER</literal>.
What about postmaster?
For parallel workers launched for parallel queries, should they follow
the backend's log level or the background worker's? Since they operate
as part of a parallel query executed by a backend, it seems more logical
for them to follow the backend's setting.
+ [B_CHECKPOINTER] = "checkpointer",
+ [B_STARTUP] = "backend", /* XXX same as backend? */I like the idea of allowing log levels to be set per process.
There were times I wanted to use debug5 specifically for
the startup process when troubleshooting WAL replay. It would be
helpful to distinguish the startup process from a regular backend,
so we can set its log level independently.
Regards,
--
Fujii Masao
Advanced Computing Technology Center
Research and Development Headquarters
NTT DATA CORPORATION
Hi,
On 2025-03-04 Tu 7:33 PM, Euler Taveira wrote:
I think it should be acceptable to configure one global setting with
exceptions for particular backend types:log_min_messages = WARNING, autovacuum:DEBUG1
Right now I think the code only accepts the unadorned log level if
there
are no other items in the list. I think the proposal downthread is to
use the keyword ALL for this,log_min_messages = all:WARNING, autovacuum:DEBUG1 # I don't
like thisbut I think it's inferior, because then "all" is not really "all",
and I
think it would be different if I had saidlog_min_messages = autovacuum:DEBUG1, all:WARNING # I don't
like thisbecause it looks like the "all" entry should override the one I set
for
autovacuum before, which frankly would not make sense to me.Good point. After reflection, I agree that "all" is not a good keyword.
This patch turns backend type as optional so WARNING means apply this
log level as a final step to the backend types that are not
specified in
the list.So I think these two lines,
log_min_messages = WARNING, autovacuum:DEBUG1
log_min_messages = autovacuum:DEBUG1, WARNINGshould behave identically and mean "set the level for autovacuum to
DEBUG1, and to any other backend type to WARNING.Done.
Just bikeshedding a bit ...
I'm not mad keen on this design. I think the value should be either a
single setting like "WARNING" or a set of type:setting pairs. I agree
that "all" is a bad name, but I think "default" would make sense.I can live with it but I think this just looks a bit odd.
Just bringing some thoughts about it...
How about using something like *:WARNING? I'm not sure if it could also be
confusing as the "all" keyword, but I think it could also be interpreted as
"anything else use WARNING".
--
Matheus Alcantara
Hi,
On 2025-03-04 21:33:39 -0300, Euler Taveira wrote:
+/* + * This must match enum BackendType! It should be static, but + * commands/variable.c needs to get at this. + */ +int log_min_messages[] = { + [B_INVALID] = WARNING, + [B_BACKEND] = WARNING, + [B_DEAD_END_BACKEND] = WARNING, + [B_AUTOVAC_LAUNCHER] = WARNING, + [B_AUTOVAC_WORKER] = WARNING, + [B_BG_WORKER] = WARNING, + [B_WAL_SENDER] = WARNING, + [B_SLOTSYNC_WORKER] = WARNING, + [B_STANDALONE_BACKEND] = WARNING, + [B_ARCHIVER] = WARNING, + [B_BG_WRITER] = WARNING, + [B_CHECKPOINTER] = WARNING, + [B_STARTUP] = WARNING, + [B_WAL_RECEIVER] = WARNING, + [B_WAL_SUMMARIZER] = WARNING, + [B_WAL_WRITER] = WARNING, + [B_LOGGER] = WARNING, +};
+StaticAssertDecl(lengthof(log_min_messages) == BACKEND_NUM_TYPES, + "array length mismatch"); + +/* + * This must match enum BackendType! It might be in commands/variable.c but for + * convenience it is near log_min_messages. + */ +const char *const log_min_messages_backend_types[] = { + [B_INVALID] = "backend", /* XXX same as backend? */ + [B_BACKEND] = "backend", + [B_DEAD_END_BACKEND] = "backend", /* XXX same as backend? */ + [B_AUTOVAC_LAUNCHER] = "autovacuum", + [B_AUTOVAC_WORKER] = "autovacuum", + [B_BG_WORKER] = "bgworker", + [B_WAL_SENDER] = "walsender", + [B_SLOTSYNC_WORKER] = "slotsyncworker", + [B_STANDALONE_BACKEND] = "backend", /* XXX same as backend? */ + [B_ARCHIVER] = "archiver", + [B_BG_WRITER] = "bgwriter", + [B_CHECKPOINTER] = "checkpointer", + [B_STARTUP] = "backend", /* XXX same as backend? */
Huh, the startup process is among the most crucial things to monitor?
+ [B_WAL_RECEIVER] = "walreceiver", + [B_WAL_SUMMARIZER] = "walsummarizer", + [B_WAL_WRITER] = "walwriter", + [B_LOGGER] = "logger", +}; + +StaticAssertDecl(lengthof(log_min_messages_backend_types) == BACKEND_NUM_TYPES, + "array length mismatch"); +
I don't know what I think about the whole patch, but I do want to voice
*strong* opposition to duplicating a list of all backend types into multiple
places. It's already painfull enough to add a new backend type, without having
to pointlessly go around and manually add a new backend type to mulltiple
arrays that have completely predictable content.
Greetings,
Andres Freund
On Wed, Mar 5, 2025, at 11:53 PM, Fujii Masao wrote:
On 2025/03/05 9:33, Euler Taveira wrote:
+ Valid <literal>BACKENDTYPE</literal> values are <literal>ARCHIVER</literal>, + <literal>AUTOVACUUM</literal>, <literal>BACKEND</literal>, + <literal>BGWORKER</literal>, <literal>BGWRITER</literal>, + <literal>CHECKPOINTER</literal>, <literal>LOGGER</literal>, + <literal>SLOTSYNCWORKER</literal>, <literal>WALRECEIVER</literal>, + <literal>WALSENDER</literal>, <literal>WALSUMMARIZER</literal>, and + <literal>WALWRITER</literal>.What about postmaster?
It is B_INVALID that is currently mapped to "backend". This state is used as an
initial phase for the child process. There might be messages before
MyBackendType is assigned (see ProcessStartupPacket, for example). Hence, I
refrain to create a special backend type for postmaster. Should we map it to
generic log level instead of backend?
For parallel workers launched for parallel queries, should they follow
the backend's log level or the background worker's? Since they operate
as part of a parallel query executed by a backend, it seems more logical
for them to follow the backend's setting.
Since we are using enum BackendType and there is no special entry for parallel
query processes. I'm afraid that introducing conditional logic for checking
special cases like the bgw_type for parallel queries or MyProcPid ==
PostmasterPid might slowdown that code path. (See that should_output_to_server
is an inline function.) I will run some tests to see if there is a considerable
impact.
+ [B_CHECKPOINTER] = "checkpointer", + [B_STARTUP] = "backend", /* XXX same as backend? */I like the idea of allowing log levels to be set per process.
There were times I wanted to use debug5 specifically for
the startup process when troubleshooting WAL replay. It would be
helpful to distinguish the startup process from a regular backend,
so we can set its log level independently.
Although startup process is mapped to backend (hence, the XXX), we can
certainly create a separate backend type for it.
--
Euler Taveira
EDB https://www.enterprisedb.com/
On Wed, Mar 5, 2025, at 1:40 PM, Andrew Dunstan wrote:
Just bikeshedding a bit ...
I'm not mad keen on this design. I think the value should be either a single setting like "WARNING" or a set of type:setting pairs. I agree that "all" is a bad name, but I think "default" would make sense.
One of my main concerns was a clear interface. I think "default" is less
confusing than "all". Your suggestion about single or list is aligned with what
Alvaro suggested. IIUC you are suggesting default:loglevel only if it is part
of the list; the single loglevel shouldn't contain the backend type to keep the
backward compatibility. The advantage of your proposal is that it make it clear
what the fallback log level is. However, someone could be confused asking if
the "default" is a valid backend type and if there is a difference between
WARNING and default:WARNING (both is a fallback for non-specified backend type
elements).
--
Euler Taveira
EDB https://www.enterprisedb.com/
On Thu, Mar 6, 2025, at 10:33 AM, Andres Freund wrote:
Huh, the startup process is among the most crucial things to monitor?
Good point. Fixed.
After collecting some suggestions, I'm attaching a new patch contains the
following changes:
- patch was rebased
- include Alvaro's patch (v2-0001) [1]/messages/by-id/202507282113.vdp4axehoppi@alvherre.pgsql as a basis for this patch
- add ioworker as new backend type
- add startup as new backend type per Andres suggestion
- small changes into documentation
I don't know what I think about the whole patch, but I do want to voice
*strong* opposition to duplicating a list of all backend types into multiple
places. It's already painfull enough to add a new backend type, without having
to pointlessly go around and manually add a new backend type to mulltiple
arrays that have completely predictable content.
I'm including Alvaro's patch as is just to make the CF bot happy and to
illustrate how it would be if we adopt his solution to centralize the list of
backend types. I think Alvaro's proposal overcomes the objection [2]/messages/by-id/y5tgui75jrcj6mm5nmoq4yqwage2432akx4kp2ogtcnim3wskx@2ipmtfi4qvpi, right?
[1]: /messages/by-id/202507282113.vdp4axehoppi@alvherre.pgsql
[2]: /messages/by-id/y5tgui75jrcj6mm5nmoq4yqwage2432akx4kp2ogtcnim3wskx@2ipmtfi4qvpi
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On Thu, 31 Jul 2025 at 11:19, "Euler Taveira" <euler@eulerto.com> wrote:
On Thu, Mar 6, 2025, at 10:33 AM, Andres Freund wrote:
Huh, the startup process is among the most crucial things to monitor?
Good point. Fixed.
After collecting some suggestions, I'm attaching a new patch contains the
following changes:- patch was rebased
- include Alvaro's patch (v2-0001) [1] as a basis for this patch
- add ioworker as new backend type
- add startup as new backend type per Andres suggestion
- small changes into documentationI don't know what I think about the whole patch, but I do want to voice
*strong* opposition to duplicating a list of all backend types into multiple
places. It's already painfull enough to add a new backend type, without having
to pointlessly go around and manually add a new backend type to mulltiple
arrays that have completely predictable content.I'm including Alvaro's patch as is just to make the CF bot happy and to
illustrate how it would be if we adopt his solution to centralize the list of
backend types. I think Alvaro's proposal overcomes the objection [2], right?
If we set the log level for all backend types, I don't think a generic log
level is necessary.
--
Regards,
Japin Li
ChengDu WenWu Information Technology Co., Ltd.
On 2025-Aug-01, Japin Li wrote:
If we set the log level for all backend types, I don't think a generic log
level is necessary.
I don't understand what you mean by this. Can you elaborate?
--
Ãlvaro Herrera 48°01'N 7°57'E â https://www.EnterpriseDB.com/
"If it is not right, do not do it.
If it is not true, do not say it." (Marcus Aurelius, Meditations)
On Thu, 31 Jul 2025 at 18:51, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
On 2025-Aug-01, Japin Li wrote:
If we set the log level for all backend types, I don't think a generic log
level is necessary.I don't understand what you mean by this. Can you elaborate?
For example:
ALTER SYSTEM SET log_min_messages TO
'archiver:info, autovacuum:info, backend:info, bgworker:info, bgwriter:info, checkpointer:info, ioworker:info, logger:info, slotsyncworker:info, startup:info, walreceiver:info, walsender:info, walsummarizer:info, walwriter:info';
Given that we've set a log level for every backend type and
assigned[BACKEND_NUM_TYPES] is true, a generic level seems redundant.
--
Regards,
Japin Li
ChengDu WenWu Information Technology Co., Ltd.
On Thu, Jul 31, 2025 at 11:19:48AM -0300, Euler Taveira wrote:
On Thu, Mar 6, 2025, at 10:33 AM, Andres Freund wrote:
Huh, the startup process is among the most crucial things to monitor?
Good point. Fixed.
After collecting some suggestions, I'm attaching a new patch contains the
following changes:- patch was rebased
- include Alvaro's patch (v2-0001) [1] as a basis for this patch
- add ioworker as new backend type
- add startup as new backend type per Andres suggestion
- small changes into documentationI don't know what I think about the whole patch, but I do want to voice
*strong* opposition to duplicating a list of all backend types into multiple
places. It's already painfull enough to add a new backend type, without having
to pointlessly go around and manually add a new backend type to mulltiple
arrays that have completely predictable content.I'm including Alvaro's patch as is just to make the CF bot happy and to
illustrate how it would be if we adopt his solution to centralize the list of
backend types. I think Alvaro's proposal overcomes the objection [2], right?
I think we can avoid memory allocation by using pg_strncasecmp().
diff --git a/src/backend/commands/variable.c b/src/backend/commands/variable.c
index 5c769dd7bcc..f854b2fac93 100644
--- a/src/backend/commands/variable.c
+++ b/src/backend/commands/variable.c
@@ -1343,14 +1343,10 @@ check_log_min_messages(char **newval, void **extra, GucSource source)
}
else
{
- char *loglevel;
- char *btype;
- bool found = false;- btype = palloc((sep - tok) + 1);
- memcpy(btype, tok, sep - tok);
- btype[sep - tok] = '\0';
- loglevel = pstrdup(sep + 1);
+ char *btype = tok;
+ char *loglevel = sep + 1;
+ bool found = false; /* Is the log level valid? */
for (entry = server_message_level_options; entry && entry->name; entry++)
@@ -1377,7 +1373,7 @@ check_log_min_messages(char **newval, void **extra, GucSource source)
found = false;
for (int i = 0; i < BACKEND_NUM_TYPES; i++)
{
- if (pg_strcasecmp(log_min_messages_backend_types[i], btype) == 0)
+ if (pg_strncasecmp(log_min_messages_backend_types[i], btype, sep - tok) == 0)
{
/* Reject duplicates for a backend type. */
if (assigned[i])--
Best regards,
Japin Li
ChengDu WenWu Information Technology Co., LTD.
On Thu, Jul 31, 2025, at 8:34 PM, Japin Li wrote:
On Thu, 31 Jul 2025 at 18:51, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
On 2025-Aug-01, Japin Li wrote:
If we set the log level for all backend types, I don't think a generic log
level is necessary.I don't understand what you mean by this. Can you elaborate?
For example:
ALTER SYSTEM SET log_min_messages TO
'archiver:info, autovacuum:info, backend:info, bgworker:info,
bgwriter:info, checkpointer:info, ioworker:info, logger:info,
slotsyncworker:info, startup:info, walreceiver:info, walsender:info,
walsummarizer:info, walwriter:info';Given that we've set a log level for every backend type and
assigned[BACKEND_NUM_TYPES] is true, a generic level seems redundant.
The main reason I didn't consider this idea was that this GUC assignment will
fail as soon as a new backend type is added. Restore a dump should work
across major versions.
I think we can avoid memory allocation by using pg_strncasecmp().
loglevel does not required a memory allocation but I didn't include the btype
part because it complicates the error message that uses btype a few lines
above.
This new patch contains the following changes:
- patch was rebased
- use commit dbf8cfb4f02
- use some GUC memory allocation functions
- avoid one memory allocation (suggested by Japin Li)
- rename backend type: logger -> syslogger
- adjust tests to increase test coverage
- improve documentation and comments to reflect the current state
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On Sun, Oct 5, 2025, at 11:18 AM, Euler Taveira wrote:
This new patch contains the following changes:
- patch was rebased
- use commit dbf8cfb4f02
- use some GUC memory allocation functions
- avoid one memory allocation (suggested by Japin Li)
- rename backend type: logger -> syslogger
- adjust tests to increase test coverage
- improve documentation and comments to reflect the current state
Patch was rebased since commit fce7c73fba4 broke it. No modifications.
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
Hi Euler,
I just reviewed the patch and got a few comments.
On Nov 6, 2025, at 21:09, Euler Taveira <euler@eulerto.com> wrote:
On Sun, Oct 5, 2025, at 11:18 AM, Euler Taveira wrote:
This new patch contains the following changes:
- patch was rebased
- use commit dbf8cfb4f02
- use some GUC memory allocation functions
- avoid one memory allocation (suggested by Japin Li)
- rename backend type: logger -> syslogger
- adjust tests to increase test coverage
- improve documentation and comments to reflect the current statePatch was rebased since commit fce7c73fba4 broke it. No modifications.
--
Euler Taveira
EDB https://www.enterprisedb.com/<v5-0001-log_min_messages-per-backend-type.patch>
1 - variable.c
```
+ /* Initialize the array. */
+ memset(newlevel, WARNING, BACKEND_NUM_TYPES * sizeof(int));
```I think this statement is wrong, because memset() writes bytes but integers, so this statement will set very byte to WARNING, which should not be the intention. You will need to use a loop to initialize every element of newlevel.
2 - variable.c
```
+ /* Parse string into list of identifiers. */
+ if (!SplitGUCList(rawstring, ',', &elemlist))
+ {
+ /* syntax error in list */
+ GUC_check_errdetail("List syntax is invalid.");
+ guc_free(rawstring);
+ list_free(elemlist);
+ return false;
+ }
```Every element of elemlist points to a position of rawstring, so it’s better to list_free(elemlist) first then guc_free(rawstring).
3 - launch_backend.c
```
static inline bool
should_output_to_server(int elevel)
{
- return is_log_level_output(elevel, log_min_messages);
+ return is_log_level_output(elevel, log_min_messages[MyBackendType]);
}
```Is it possible that when this function is called, MyBackendType has not been initialized? To be safe, maybe we can check if MyBackendType is 0 (B_INVALID), then use the generic log_min_message.
4 - config.sgml
```
+ Valid values are a comma-separated list of <literal>backendtype:level</literal>
+ and a single <literal>level</literal>. The list allows it to use
+ different levels per backend type. Only the single <literal>level</literal>
+ is mandatory (order does not matter) and it is assigned to the backend
+ types that are not specified in the list.
+ Valid <literal>backendtype</literal> values are <literal>archiver</literal>,
+ <literal>autovacuum</literal>, <literal>backend</literal>,
+ <literal>bgworker</literal>, <literal>bgwriter</literal>,
+ <literal>checkpointer</literal>, <literal>ioworker</literal>,
+ <literal>syslogger</literal>, <literal>slotsyncworker</literal>,
+ <literal>startup</literal>, <literal>walreceiver</literal>,
+ <literal>walsender</literal>, <literal>walsummarizer</literal>, and
+ <literal>walwriter</literal>.
+ Valid <literal>level</literal> values are <literal>DEBUG5</literal>,
+ <literal>DEBUG4</literal>, <literal>DEBUG3</literal>, <literal>DEBUG2</literal>,
+ <literal>DEBUG1</literal>, <literal>INFO</literal>, <literal>NOTICE</literal>,
+ <literal>WARNING</literal>, <literal>ERROR</literal>, <literal>LOG</literal>,
+ <literal>FATAL</literal>, and <literal>PANIC</literal>. Each level includes
+ all the levels that follow it. The later the level, the fewer messages are sent
+ to the log. The default is <literal>WARNING</literal> for all backend types.
+ Note that <literal>LOG</literal> has a different rank here than in
```* “Valid values are …”, here “are” usually mean both are needed, so maybe change “are” to “can be”.
* It says “the single level is mandatory”, then why there is still a default value?
Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/
On 2025-Nov-06, Euler Taveira wrote:
+ Valid values are a comma-separated list of <literal>backendtype:level</literal> + and a single <literal>level</literal>. The list allows it to use + different levels per backend type. Only the single <literal>level</literal> + is mandatory (order does not matter) and it is assigned to the backend + types that are not specified in the list. + Valid <literal>backendtype</literal> values are <literal>archiver</literal>, + <literal>autovacuum</literal>, <literal>backend</literal>, + <literal>bgworker</literal>, <literal>bgwriter</literal>, + <literal>checkpointer</literal>, <literal>ioworker</literal>, + <literal>syslogger</literal>, <literal>slotsyncworker</literal>, + <literal>startup</literal>, <literal>walreceiver</literal>, + <literal>walsender</literal>, <literal>walsummarizer</literal>, and + <literal>walwriter</literal>. + Valid <literal>level</literal> values are <literal>DEBUG5</literal>, + <literal>DEBUG4</literal>, <literal>DEBUG3</literal>, <literal>DEBUG2</literal>, + <literal>DEBUG1</literal>, <literal>INFO</literal>, <literal>NOTICE</literal>, + <literal>WARNING</literal>, <literal>ERROR</literal>, <literal>LOG</literal>, + <literal>FATAL</literal>, and <literal>PANIC</literal>. Each level includes + all the levels that follow it. The later the level, the fewer messages are sent + to the log. The default is <literal>WARNING</literal> for all backend types.
I would use <literal>type:level</literal> rather than "backendtype".
Also, the glossary says we have "auxiliary processes" and "backends", so
from the user perspective, these types are not all backends, but instead
process types.
I think the list of backend types is pretty hard to read. What do you
think about using
<simplelist type="vert" columns="4">
to create a friendlier representation?
So you would say something like "Valid types are listed in the table
below, each corresponding to either postmaster, an auxiliary process
type or a backend". (Eh, wait, you seem not to have "postmaster" in
your list, what's up with that?)
(I'm not sure about making the log levels also a <simplelist>, because
for that list the order matters, and if we have more than one column
then the order is going to be ambiguous, and if we have just one column
it's going to be too tall. But maybe it's not all that bad?)
--
Ãlvaro Herrera 48°01'N 7°57'E â https://www.EnterpriseDB.com/
On Thu, Nov 6, 2025, at 1:01 PM, Alvaro Herrera wrote:
I would use <literal>type:level</literal> rather than "backendtype".
Also, the glossary says we have "auxiliary processes" and "backends", so
from the user perspective, these types are not all backends, but instead
process types.
Hasn't that ship already sailed? If your suggestion is limited to "type:level",
that's ok. I wouldn't like to use 2 terminologies ("process type" and "backend
type") in the documentation.
See miscadmin.h.
/*
* MyBackendType indicates what kind of a backend this is.
*
* If you add entries, please also update the child_process_kinds array in
* launch_backend.c.
*/
typedef enum BackendType
The "backend type" terminology is exposed. It is even available in the
pg_stat_activity view. I agree that "backend" is not a good name to define a
Postgres process. I don't think we should be inconsistent only in this patch.
Even if the proposal is to rename enum BackendType (I won't propose it as part
of this patch), it would make some extension authors unhappy according to
codesearch.debian.net.
I think the list of backend types is pretty hard to read. What do you
think about using
<simplelist type="vert" columns="4">
to create a friendlier representation?
I tried but didn't like the result. See the attached image. You said table but
it is a multi-column list; it doesn't contain a header or borders. Reading this
message and Chao Li message, I realized the description is not good enough yet.
So you would say something like "Valid types are listed in the table
below, each corresponding to either postmaster, an auxiliary process
type or a backend". (Eh, wait, you seem not to have "postmaster" in
your list, what's up with that?)
Good question. The current patch uses "backend" to B_INVALID (that's exactly the
MyBackendType for postmaster -- see below). I think it is reasonable to create a
new category "postmaster" and assign it to B_INVALID. If, for some reason, a
process temporarily has MyBackendType equals to B_INVALID, it is ok to still
consider it a postmaster process.
$ ps aux | grep postgres
euler 56968 0.0 0.0 217144 29016 ? Ss 19:28 0:00 /c/pg/bin/postgres -D /c/pg/pgdata
euler 56969 0.0 0.0 217276 9948 ? Ss 19:28 0:00 postgres: io worker 0
euler 56970 0.0 0.0 217276 7620 ? Ss 19:28 0:00 postgres: io worker 1
euler 56971 0.0 0.0 217144 6408 ? Ss 19:28 0:00 postgres: io worker 2
euler 56972 0.0 0.0 217276 8388 ? Ss 19:28 0:00 postgres: checkpointer
euler 56973 0.0 0.0 217300 8744 ? Ss 19:28 0:00 postgres: background writer
euler 56975 0.0 0.0 217276 11292 ? Ss 19:28 0:00 postgres: walwriter
euler 56976 0.0 0.0 218724 9140 ? Ss 19:28 0:00 postgres: autovacuum launcher
euler 56977 0.0 0.0 218724 9232 ? Ss 19:28 0:00 postgres: logical replication launcher
euler 58717 0.0 0.0 6676 2232 pts/1 S+ 19:39 0:00 grep --color=auto postgres
$ for p in $(pgrep postgres); do echo "pid: $p" && gdb -q -batch -ex 'set pagination off' -ex "attach $p" -ex 'p MyBackendType' -ex 'quit' 2> /dev/null; done | grep -E '^\$1|pid'
pid: 56968
$1 = B_INVALID
pid: 56969
$1 = B_IO_WORKER
pid: 56970
$1 = B_IO_WORKER
pid: 56971
$1 = B_IO_WORKER
pid: 56972
$1 = B_CHECKPOINTER
pid: 56973
$1 = B_BG_WRITER
pid: 56975
$1 = B_WAL_WRITER
pid: 56976
$1 = B_AUTOVAC_LAUNCHER
pid: 56977
$1 = B_BG_WORKER
(I'm not sure about making the log levels also a <simplelist>, because
for that list the order matters, and if we have more than one column
then the order is going to be ambiguous, and if we have just one column
it's going to be too tall. But maybe it's not all that bad?)
+1.
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On 2025-Nov-18, Euler Taveira wrote:
On Thu, Nov 6, 2025, at 1:01 PM, Alvaro Herrera wrote:
See miscadmin.h.
/*
* MyBackendType indicates what kind of a backend this is.
*
* If you add entries, please also update the child_process_kinds array in
* launch_backend.c.
*/
typedef enum BackendTypeThe "backend type" terminology is exposed.
Something appearing in the source code does not equate it being exposed
to end users. Things are exposed when they are in the documentation, or
when the SQL interface requires you to use the term. So I don't think
the fact that BackendType appears in the code forces us to use that name
in the user-visible interface now.
In the glossary, we talk about "processes"; in the definitions there,
"backend" is one type of process. So in this list of "backend types"
that you want to introduce, what you really should be talking about is
"process types", where "backend" is one of the options.
It is even available in the
pg_stat_activity view. I agree that "backend" is not a good name to define a
Postgres process. I don't think we should be inconsistent only in this patch.
Even if the proposal is to rename enum BackendType (I won't propose it as part
of this patch), it would make some extension authors unhappy according to
codesearch.debian.net.
I don't think we should rename the BackendType enum. That's in the
source code, not in the documentation or the SQL interface, so we don't
need to care.
I think the list of backend types is pretty hard to read. What do you
think about using
<simplelist type="vert" columns="4">
to create a friendlier representation?I tried but didn't like the result. See the attached image.
Well, I like it very much. The original is a solid wall of text, very
hard to read, where you have to squint hunting commas in order to
distinguish one item from the next. (Maybe you are young and have good
eyesight, but you won't be young forever.) In the screenshot you show,
the list of possible process types to use is nicely separated, which
makes it very easy to catch at a glance. Maybe remove "the table",
which is obviously inappropriate, and just say "Valid process types are
listed below, each corresponding to either postmaster, an auxiliary
process type or a backend".
(Note your original wording says "a backend type, corresponding to [blah
blah] or a backend" which makes no sense -- how is a backend a type of
backend? It isn't. It's a type of process.)
You said table but it is a multi-column list; it doesn't contain a
header or borders.
Right. You don't need a full-blown table here: this simple list is
perfectly adequate.
Good question. The current patch uses "backend" to B_INVALID (that's exactly the
MyBackendType for postmaster -- see below). I think it is reasonable to create a
new category "postmaster" and assign it to B_INVALID.
I guess that would work, but I think it's inadequate. Maybe we could
add a new value B_POSTMASTER and have postmaster switch to that as early
as possible. Then anything that still has B_INVALID must necessarily be
an improperly identified process. Users wouldn't assign a value to that
one (the GUC wouldn't let you); instead those would always use the
default value. Hopefully nobody would see that very often, or at all.
--
Ãlvaro Herrera Breisgau, Deutschland â https://www.EnterpriseDB.com/
On Wed, Nov 19, 2025, at 7:44 AM, Alvaro Herrera wrote:
On 2025-Nov-18, Euler Taveira wrote:
The "backend type" terminology is exposed.
Something appearing in the source code does not equate it being exposed
to end users. Things are exposed when they are in the documentation, or
when the SQL interface requires you to use the term. So I don't think
the fact that BackendType appears in the code forces us to use that name
in the user-visible interface now.
It is *already* exposed.
$ grep -B 3 -A 3 'backend type' doc/src/sgml/config.sgml
</informaltable>
<para>
The backend type corresponds to the column
<structfield>backend_type</structfield> in the view
<link linkend="monitoring-pg-stat-activity-view">
<structname>pg_stat_activity</structname></link>,
--
character count of the error position therein,
location of the error in the PostgreSQL source code
(if <varname>log_error_verbosity</varname> is set to <literal>verbose</literal>),
application name, backend type, process ID of parallel group leader,
and query id.
Here is a sample table definition for storing CSV-format log output:
$ grep -B 3 -A 3 'backend_type' doc/src/sgml/config.sgml
<para>
The backend type corresponds to the column
<structfield>backend_type</structfield> in the view
<link linkend="monitoring-pg-stat-activity-view">
<structname>pg_stat_activity</structname></link>,
but additional types can appear
--
query_pos integer,
location text,
application_name text,
backend_type text,
leader_pid integer,
query_id bigint,
PRIMARY KEY (session_id, session_line_num)
--
<entry>Client application name</entry>
</row>
<row>
<entry><literal>backend_type</literal></entry>
<entry>string</entry>
<entry>Type of backend</entry>
</row>
$ psql -c "SELECT backend_type FROM pg_stat_activity" -d postgres
backend_type
------------------------------
client backend
autovacuum launcher
logical replication launcher
io worker
io worker
io worker
checkpointer
background writer
walwriter
(9 rows)
In the glossary, we talk about "processes"; in the definitions there,
"backend" is one type of process. So in this list of "backend types"
that you want to introduce, what you really should be talking about is
"process types", where "backend" is one of the options.It is even available in the
pg_stat_activity view. I agree that "backend" is not a good name to define a
Postgres process. I don't think we should be inconsistent only in this patch.
Even if the proposal is to rename enum BackendType (I won't propose it as part
of this patch), it would make some extension authors unhappy according to
codesearch.debian.net.I don't think we should rename the BackendType enum. That's in the
source code, not in the documentation or the SQL interface, so we don't
need to care.
I think you missed my point here. As I said if your proposal will let us with 2
terminologies (backend type and process type) as shown above. We can debate if
we (1) keep the current terminology (backend type), (2) use the proposed one
(process type) or (3) use a mixed variation (keep the SQL interface --
backend_type column but change the description to "process type"). I wouldn't
like to break compatibility (pg_stat_activity changes tend to break a lot of
stuff) so I'm fine with (1) and (3).
I think the list of backend types is pretty hard to read. What do you
think about using
<simplelist type="vert" columns="4">
to create a friendlier representation?I tried but didn't like the result. See the attached image.
Well, I like it very much. The original is a solid wall of text, very
hard to read, where you have to squint hunting commas in order to
distinguish one item from the next. (Maybe you are young and have good
eyesight, but you won't be young forever.) In the screenshot you show,
the list of possible process types to use is nicely separated, which
makes it very easy to catch at a glance. Maybe remove "the table",
which is obviously inappropriate, and just say "Valid process types are
listed below, each corresponding to either postmaster, an auxiliary
process type or a backend".
I don't have a strong preference. One advantage of this suggestion is that it
is visually easier to identify the types.
(Note your original wording says "a backend type, corresponding to [blah
blah] or a backend" which makes no sense -- how is a backend a type of
backend? It isn't. It's a type of process.)
I'm rewriting that paragraph.
Good question. The current patch uses "backend" to B_INVALID (that's exactly the
MyBackendType for postmaster -- see below). I think it is reasonable to create a
new category "postmaster" and assign it to B_INVALID.I guess that would work, but I think it's inadequate. Maybe we could
add a new value B_POSTMASTER and have postmaster switch to that as early
as possible. Then anything that still has B_INVALID must necessarily be
an improperly identified process. Users wouldn't assign a value to that
one (the GUC wouldn't let you); instead those would always use the
default value. Hopefully nobody would see that very often, or at all.
I will create a separate patch for this suggestion.
--
Euler Taveira
EDB https://www.enterprisedb.com/
On Fri, Nov 21, 2025, at 11:13 AM, Euler Taveira wrote:
I think you missed my point here. As I said if your proposal will let us with 2
terminologies (backend type and process type) as shown above. We can debate if
we (1) keep the current terminology (backend type), (2) use the proposed one
(process type) or (3) use a mixed variation (keep the SQL interface --
backend_type column but change the description to "process type"). I wouldn't
like to break compatibility (pg_stat_activity changes tend to break a lot of
stuff) so I'm fine with (1) and (3).
After digesting it for a couple of days, I decided to adopt the terminology in
commit dbf8cfb4f02e (process type) just for this patch. This discussion about
changing "backend type" terminology in other parts belongs to another patch.
Good question. The current patch uses "backend" to B_INVALID (that's exactly the
MyBackendType for postmaster -- see below). I think it is reasonable to create a
new category "postmaster" and assign it to B_INVALID.I guess that would work, but I think it's inadequate. Maybe we could
add a new value B_POSTMASTER and have postmaster switch to that as early
as possible. Then anything that still has B_INVALID must necessarily be
an improperly identified process. Users wouldn't assign a value to that
one (the GUC wouldn't let you); instead those would always use the
default value. Hopefully nobody would see that very often, or at all.I will create a separate patch for this suggestion.
After reflection, add B_POSTMASTER into BackendType / proctypelist.h is not a
good idea. These data structures were designed to represent postmaster
children. The child_process_kinds array (launch_backend.c) is assembled with
proctypelist.h which means a function like postmaster_child_launch() could use
one of its elements (e.g. B_POSTMASTER) to start a new child; that's awkward. Do
you envision any issues to use B_INVALID for postmaster? (I wrote 0002 to
minimize the windows that each child process has B_INVALID.)
This new version contains the following changes:
- fix variable assignment (Chao Li)
- fix memory release order (Chao Li)
- change category for B_INVALID (backend -> postmaster) (Alvaro)
- rewrite documentation (Chao Li, Alvaro)
- rename backend type to process type (Alvaro)
- new patch (0002) that assigns MyBackendType as earlier as possible
--
Euler Taveira
EDB https://www.enterprisedb.com/
Attachments:
On Thu, Nov 6, 2025, at 11:53 AM, Chao Li wrote:
I just reviewed the patch and got a few comments.
Thanks for your review.
+ /* Initialize the array. */ + memset(newlevel, WARNING, BACKEND_NUM_TYPES * sizeof(int)); ```I think this statement is wrong, because memset() writes bytes but
integers, so this statement will set very byte to WARNING, which should
not be the intention. You will need to use a loop to initialize every
element of newlevel.
Facepalm. Good catch.
2 - variable.c ``` + /* Parse string into list of identifiers. */ + if (!SplitGUCList(rawstring, ',', &elemlist)) + { + /* syntax error in list */ + GUC_check_errdetail("List syntax is invalid."); + guc_free(rawstring); + list_free(elemlist); + return false; + } ```Every element of elemlist points to a position of rawstring, so it’s
better to list_free(elemlist) first then guc_free(rawstring).
Fixed.
3 - launch_backend.c ``` static inline bool should_output_to_server(int elevel) { - return is_log_level_output(elevel, log_min_messages); + return is_log_level_output(elevel, log_min_messages[MyBackendType]); } ```Is it possible that when this function is called, MyBackendType has not
been initialized? To be safe, maybe we can check if MyBackendType is 0
(B_INVALID), then use the generic log_min_message.
It uses the generic log level if you don't modify backend type "backend". If
you do specify "backend", that level is used instead. After Alvaro's question
in the other email, I added "postmaster" backend type and assigned it to
B_INVALID. That's exactly the log level that will be used. As I said in that
email, it is ok to consider a process whose backend type is B_INVALID as a
postmaster process.
* “Valid values are …”, here “are” usually mean both are needed, so
maybe change “are” to “can be”.
* It says “the single level is mandatory”, then why there is still a
default value?
It seems the current sentence that describe the comma-separated list is
ambiguous. The sentence doesn't make it clear that the list contains 2 elements
(type:level and level) and the "level" element is mandatory. What about the new
sentence?
--
Euler Taveira
EDB https://www.enterprisedb.com/
Hello
I noticed failures under Windows, and realized that the EXEC_BACKEND
case was busted. That's easy to fix -- just assign MyBackendType in
SubPostmasterMain() as well. With that fix I was still getting some
crashes in three test modules, and after some debugging it turned out
that if you have shared_preload_library with a background worker and use
%b in the log_line_prefix, you get a crash trying to expand a name that
hasn't been set up yet. I figured we can use a constant string in that
case. Better ideas for that string welcome.
So here's your v6 again with those fixes as 0003 -- let's see what CI
thinks of this. I haven't looked at your doc changes yet.
BTW with %b in log_line_prefix, the log file looks ... interesting.
--
Álvaro Herrera Breisgau, Deutschland — https://www.EnterpriseDB.com/
"Before you were born your parents weren't as boring as they are now. They
got that way paying your bills, cleaning up your room and listening to you
tell them how idealistic you are." -- Charles J. Sykes' advice to teenagers
Attachments:
BTW another thing I realized while looking this over, is that we quite
uselessly transform the integer backend type to a string, pass it as a
string using the --forkchild= argument to the child process, then parse
the string back to an int to use as an array index. It would be much
easier to just use the integer value everywhere, as the attached shows.
--
Álvaro Herrera 48°01'N 7°57'E — https://www.EnterpriseDB.com/
"I love the Postgres community. It's all about doing things _properly_. :-)"
(David Garamond)
Attachments:
On 2025-Dec-09, Alvaro Herrera wrote:
So here's your v6 again with those fixes as 0003 -- let's see what CI
thinks of this. I haven't looked at your doc changes yet.
This passed CI, so I have marked it as Ready for Committer. Further
comments are still welcome, of course, but if there are none, I intend
to get it committed in a few days.
I'm not really happy with our usage of the translatable description
field for things such as %b in log_line_prefix or the ps display; I
think we should use the shorter names there instead. Otherwise, you end
up with log lines that look something like this (visible in a server
with %b in log_line_prefix, which the TAP tests as well as pg_regress
have):
2025-12-08 21:38:04.304 CET autovacuum launcher[2452437] DEBUG: autovacuum launcher started
where the bit before the PID is marked for translation. I think it
should rather be
2025-12-08 21:38:04.304 CET autovacuum[2452437] DEBUG: autovacuum launcher started
where that name (the same we'll use in log_min_messages) is not
translated.
However, this issue is rather independent of the patch in this thread,
so I'm going to discuss it in another thread; the name string though is
added by this patch.
--
Álvaro Herrera Breisgau, Deutschland — https://www.EnterpriseDB.com/
"Java is clearly an example of money oriented programming" (A. Stepanov)
On Dec 10, 2025, at 02:24, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
On 2025-Dec-09, Alvaro Herrera wrote:
So here's your v6 again with those fixes as 0003 -- let's see what CI
thinks of this. I haven't looked at your doc changes yet.This passed CI, so I have marked it as Ready for Committer. Further
comments are still welcome, of course, but if there are none, I intend
to get it committed in a few days.I'm not really happy with our usage of the translatable description
field for things such as %b in log_line_prefix or the ps display; I
think we should use the shorter names there instead. Otherwise, you end
up with log lines that look something like this (visible in a server
with %b in log_line_prefix, which the TAP tests as well as pg_regress
have):2025-12-08 21:38:04.304 CET autovacuum launcher[2452437] DEBUG: autovacuum launcher started
where the bit before the PID is marked for translation. I think it
should rather be2025-12-08 21:38:04.304 CET autovacuum[2452437] DEBUG: autovacuum launcher started
where that name (the same we'll use in log_min_messages) is not
translated.However, this issue is rather independent of the patch in this thread,
so I'm going to discuss it in another thread; the name string though is
added by this patch.--
Álvaro Herrera Breisgau, Deutschland — https://www.EnterpriseDB.com/
"Java is clearly an example of money oriented programming" (A. Stepanov)
Overall the patch is solid, I just have a couple of suggestions:
1 - user experience
```
evantest=# show log_min_messages;
log_min_messages
---------------------
warning,backend:log
(1 row)
```
Now “show log_min_messages” prints the raw string the user set, in above example, there is not a white-space between the two log levels, and “show” result doesn’t have a white-space between the two log levels either. IMO, “SHOW log_min_messages” should display a stable result, in other words, say “fatal, backend:log” and “backend:log, fatal” should show the same result as they are actually meaning the same. So, I would suggest normalize the raw string: put the general level in the first place and sort others by process type, then SHOW returns the normalized string.
2 - refactoring for 0001
```
+ sep = strchr(tok, ':');
+ if (sep == NULL)
+ {
+ bool found = false;
+
+ /* Reject duplicates for generic log level. */
+ if (genericlevel != -1)
+ {
+ GUC_check_errdetail("Generic log level was already assigned.");
+ guc_free(rawstring);
+ list_free(elemlist);
+ return false;
+ }
+
+ /* Is the log level valid? */
+ for (entry = server_message_level_options; entry && entry->name; entry++)
+ {
+ if (pg_strcasecmp(entry->name, tok) == 0)
+ {
+ genericlevel = entry->val;
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ GUC_check_errdetail("Unrecognized log level: \"%s\".", tok);
+ guc_free(rawstring);
+ list_free(elemlist);
+ return false;
+ }
+ }
+ else
+ {
+ char *loglevel;
+ char *ptype;
+ bool found = false;
+
+ ptype = guc_malloc(LOG, (sep - tok) + 1);
+ if (!ptype)
+ {
+ guc_free(rawstring);
+ list_free(elemlist);
+ return false;
+ }
+ memcpy(ptype, tok, sep - tok);
+ ptype[sep - tok] = '\0';
+ loglevel = sep + 1;
+
+ /* Is the log level valid? */
+ for (entry = server_message_level_options; entry && entry->name; entry++)
+ {
+ if (pg_strcasecmp(entry->name, loglevel) == 0)
+ {
+ found = true;
+ break;
+ }
+ }
+
+ if (!found)
+ {
+ GUC_check_errdetail("Unrecognized log level: \"%s\".", loglevel);
+ guc_free(ptype);
+ guc_free(rawstring);
+ list_free(elemlist);
+ return false;
+ }
```In the “if” and “else” clauses, there are duplicate code to valid log levels. We should refactor the code to avoid the duplication. For example, pull up “loglevel” to the “for” loop level, then we can valid it after the “if-else”.
Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/
On Tue, Dec 9, 2025, at 11:00 PM, Chao Li wrote:
Now “show log_min_messages” prints the raw string the user set, in
above example, there is not a white-space between the two log levels,
and “show” result doesn’t have a white-space between the two log levels
either. IMO, “SHOW log_min_messages” should display a stable result, in
other words, say “fatal, backend:log” and “backend:log, fatal” should
show the same result as they are actually meaning the same. So, I would
suggest normalize the raw string: put the general level in the first
place and sort others by process type, then SHOW returns the normalized
string.
I thought about it but leave it alone because (a) it would increase this patch
footprint and (b) the input might be different from the output. I could also be
done in another patch but under reflection an unstable output can break tests
or whatever uses the SHOW log_min_messages output. I thought this change would
require a new show_log_min_messages to manipulate the input again but we can
reassign the GUC value after sorting the existing list and creating a new string
list.
In the “if” and “else” clauses, there are duplicate code to valid log
levels. We should refactor the code to avoid the duplication. For
example, pull up “loglevel” to the “for” loop level, then we can valid
it after the “if-else”.
The for loop is duplicate but if you create a separate function for it but the
result is:
src/backend/commands/variable.c | 43 ++++++++++++++++++++++---------------------
1 file changed, 22 insertions(+), 21 deletions(-)
I'll post a patch in a couple of hours after spend more time in it.
--
Euler Taveira
EDB https://www.enterprisedb.com/
On Dec 11, 2025, at 09:57, Euler Taveira <euler@eulerto.com> wrote:
In the “if” and “else” clauses, there are duplicate code to valid log
levels. We should refactor the code to avoid the duplication. For
example, pull up “loglevel” to the “for” loop level, then we can valid
it after the “if-else”.The for loop is duplicate but if you create a separate function for it but the
result is:src/backend/commands/variable.c | 43 ++++++++++++++++++++++---------------------
1 file changed, 22 insertions(+), 21 deletions(-)
I don’t think we need to add a separate function. We can use ‘if-else” to parse log level, then verify it after “if-else”.
Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/