Copilot commented on code in PR #13741:
URL: https://github.com/apache/ignite/pull/13741#discussion_r4376418937


##########
docs/_docs/logging.adoc:
##########
@@ -35,7 +35,7 @@ When a node starts, it outputs start-up information to the 
console, including th
 | `IGNITE_LOG_INSTANCE_NAME` | If the property is set, Ignite includes its 
instance name in log messages. |  Not set
 | `IGNITE_QUIET` | Set to `false` to disable the quiet mode and enable the 
verbose mode.
 In the verbose mode, the node logs a lot more information. | `true`
-| `IGNITE_LOG_DIR` | The directory where Ignite writes log files. | 
`$IGNITE_HOME/
+| `IGNITE_LOG_DIR` | The directory where the built-in JUL logger writes log 
files. Set it as an environment variable of the node process; it does not 
affect Log4j2 appenders. | `$IGNITE_HOME/
 work/log`

Review Comment:
   `IGNITE_LOG_DIR` is commonly configured via a JVM system property 
(`-DIGNITE_LOG_DIR=...`) even when it’s sourced from an environment variable by 
wrapper scripts (e.g., `ignite.sh`). The updated text states it must be set as 
an environment variable, which may be inaccurate/misleading for users launching 
Ignite directly (systemd/K8s/IDE) and setting JVM opts. Recommendation: 
describe it as a JVM system property and optionally mention that distribution 
scripts can populate it from an environment variable; keep the note that it 
doesn’t affect Log4j2 appenders.



##########
docs/_docs/logging.adoc:
##########
@@ -44,7 +44,12 @@ work/log`
 == Default Logging
 By default, Ignite uses the java.util.logging (JUL) framework.
 If you start Ignite using the `ignite.sh|bat` script from the distribution 
package, Ignite uses `$IGNITE_HOME/config/java.util.logging.properties` as the 
default logging configuration file and outputs all messages to log files in the 
`$IGNITE_HOME/work/log` directory.
-You can override the default logging directory by specifying the 
`IGNITE_LOG_DIR` system property.
+You can override the default logging directory by setting the `IGNITE_LOG_DIR` 
environment variable for the node process. This variable is read once at 
startup and applies to the built-in JUL logger only.

Review Comment:
   `IGNITE_LOG_DIR` is commonly configured via a JVM system property 
(`-DIGNITE_LOG_DIR=...`) even when it’s sourced from an environment variable by 
wrapper scripts (e.g., `ignite.sh`). The updated text states it must be set as 
an environment variable, which may be inaccurate/misleading for users launching 
Ignite directly (systemd/K8s/IDE) and setting JVM opts. Recommendation: 
describe it as a JVM system property and optionally mention that distribution 
scripts can populate it from an environment variable; keep the note that it 
doesn’t affect Log4j2 appenders.



##########
docs/_docs/persistence/persistence-tuning.adoc:
##########
@@ -63,6 +63,9 @@ tab:C++[unsupported]
 
 Consider using separate drives for data files and 
link:persistence/native-persistence#write-ahead-log[Write-Ahead-Logging (WAL)].
 Ignite actively writes to both the data and WAL files.
+Keeping the active WAL on its own device reduces the competition between 
checkpoint page writes and WAL writes, which helps keep write latency stable 
while a checkpoint is in progress.
+The durability guarantees differ by WAL mode: with the default `LOG_ONLY` mode 
the WAL is flushed to the operating system cache rather than forced to disk on 
every commit, while `FSYNC` waits for the disk.
+See link:server-node-file-system-layout[Server Node File System Layout] for 
the complete placement of data, WAL, the WAL archive, CDC, snapshots, logs, and 
diagnostic files.

Review Comment:
   This paragraph is duplicated verbatim in 
`docs/_docs/perf-and-troubleshooting/persistence-tuning.adoc` (same change in 
this PR). Duplicating operational guidance across multiple pages increases the 
chance of future drift. If both pages are intended to stay aligned, consider 
extracting this shared text into an `include::` snippet used by both pages 
(optional, but helps keep future edits consistent).



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to