Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Thu, 6 Apr 2017 01:35:41 +0000 (UTC)
From: "Nemo Chen (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12999039.1471904686000.219101.1491442541595@Atlassian.JIRA>
In-Reply-To: <JIRA.12999039.1471904686000@Atlassian.JIRA>
References: <JIRA.12999039.1471904686000@Atlassian.JIRA> <JIRA.12999039.1471904686333@jira-lw-us.apache.org>
Subject: [jira] [Updated] (HBASE-16469) Several log refactoring/improvement
 suggestions
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Thu, 06 Apr 2017 01:35:50 -0000


     [ https://issues.apache.org/jira/browse/HBASE-16469?page=3Dcom.atlassi=
an.jira.plugin.system.issuetabpanels:all-tabpanel ]

Nemo Chen updated HBASE-16469:
------------------------------
    Description:=20
*method invocation replaced by variable*

hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/handler/Clo=
seRegionHandler.java

line 118: {code}String name =3D regionInfo.getRegionNameAsString();{code}
line 142: {code}LOG.warn("Can't close region: was already closed during clo=
se(): " +
            regionInfo.getRegionNameAsString()); {code}

In the above two examples, the method invocations are assigned to the varia=
bles before the logging code. These method invocations should be replaced b=
y variables in case of simplicity and readability
----

*method invocation in return statement*

hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.jav=
a

line 5455:
{code}
public String toString() {
    return getRegionInfo().getRegionNameAsString();
  }
{code}

line 1260:
{code}
LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
          + " is not mergeable because it is closing or closed");
{code}
line 1265:
{code}
LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
          + " is not mergeable because it has references");
{code}
line 1413:
{code}=20
LOG.info("Running close preflush of " + getRegionInfo().getRegionNameAsStri=
ng());
{code}

In these above examples, the "getRegionInfo().getRegionNameAsString())" is =
the return statement of method "toString" in the same class. They should be=
 replaced with =E2=80=9Cthis=E2=80=9D   in case of simplicity and readabili=
ty.

----
*check the logged variable if it is null*
hbase-it/src/test/java/org/apache/hadoop/hbase/HBaseClusterManager.java

line 88:=20
{code}
if ((sshUserName !=3D null && sshUserName.length() > 0) ||
        (sshOptions !=3D null && sshOptions.length() > 0)) {
      LOG.info("Running with SSH user [" + sshUserName + "] and options [" =
+ sshOptions + "]");
    }
{code}

hbase-server/src/main/java/org/apache/hadoop/hbase/master/AssignmentManager=
.java

line 980:
{code}
if ((regionState =3D=3D null && latestState !=3D null)
          || (regionState !=3D null && latestState =3D=3D null)
          || (regionState !=3D null && latestState !=3D null
            && latestState.getState() !=3D regionState.getState())) {
        LOG.warn("Region state changed from " + regionState + " to "
          + latestState + ", while acquiring lock");
      }
{code}
In the above example, the logging variable could null at run time. It is a =
bad  practice to include null variables inside logs.

----
*variable in byte printed directly*

hbase-server/src/test/java/org/apache/hadoop/hbase/util/MultiThreadedUpdate=
r.java

line 145:=20
{code}
byte[] rowKey =3D dataGenerator.getDeterministicUniqueKey(rowKeyBase);
{code}
line 184:
{code}
LOG.error("Failed to update the row with key =3D [" + rowKey
      + "], since we could not get the original row");
{code}

rowKey should be printed as Bytes.toString(rowKey).

----=20

*object toString contain mi*

The toString method returns getServerName(), so the "server.getServerName()=
" should be replaced with "server" in case of simplicity and readability.

hbase-client/src/main/java/org/apache/hadoop/hbase/client/PreemptiveFastFai=
lInterceptor.java

{code}
LOG.info("Clearing out PFFE for server " + server.getServerName());
return getServerName();
{code}


  was:
*method invocation replaced by variable*

hbase-1.2.2/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver=
/handler/CloseRegionHandler.java

line 118: {code}String name =3D regionInfo.getRegionNameAsString();{code}
line 142: {code}LOG.warn("Can't close region: was already closed during clo=
se(): " +
            regionInfo.getRegionNameAsString()); {code}

In the above two examples, the method invocations are assigned to the varia=
bles before the logging code. These method invocations should be replaced b=
y variables in case of simplicity and readability
----

*method invocation in return statement*

hbase-1.2.2/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver=
/HRegion.java

line 5455:
{code}
public String toString() {
    return getRegionInfo().getRegionNameAsString();
  }
{code}

line 1260:
{code}
LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
          + " is not mergeable because it is closing or closed");
{code}
line 1265:
{code}
LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
          + " is not mergeable because it has references");
{code}
line 1413:
{code}=20
LOG.info("Running close preflush of " + getRegionInfo().getRegionNameAsStri=
ng());
{code}

In these above examples, the "getRegionInfo().getRegionNameAsString())" is =
the return statement of method "toString" in the same class. They should be=
 replaced with =E2=80=9Cthis=E2=80=9D   in case of simplicity and readabili=
ty.

----
*check the logged variable if it is null*
hbase-1.2.2/hbase-it/src/test/java/org/apache/hadoop/hbase/HBaseClusterMana=
ger.java

line 88:=20
{code}
if ((sshUserName !=3D null && sshUserName.length() > 0) ||
        (sshOptions !=3D null && sshOptions.length() > 0)) {
      LOG.info("Running with SSH user [" + sshUserName + "] and options [" =
+ sshOptions + "]");
    }
{code}

hbase-1.2.2/hbase-server/src/main/java/org/apache/hadoop/hbase/master/Assig=
nmentManager.java

line 980:
{code}
if ((regionState =3D=3D null && latestState !=3D null)
          || (regionState !=3D null && latestState =3D=3D null)
          || (regionState !=3D null && latestState !=3D null
            && latestState.getState() !=3D regionState.getState())) {
        LOG.warn("Region state changed from " + regionState + " to "
          + latestState + ", while acquiring lock");
      }
{code}
In the above example, the logging variable could null at run time. It is a =
bad  practice to include null variables inside logs.

----
*variable in byte printed directly*

hbase-1.2.2/hbase-server/src/test/java/org/apache/hadoop/hbase/util/MultiTh=
readedUpdater.java

line 145:=20
{code}
byte[] rowKey =3D dataGenerator.getDeterministicUniqueKey(rowKeyBase);
{code}
line 184:
{code}
LOG.error("Failed to update the row with key =3D [" + rowKey
      + "], since we could not get the original row");
{code}

rowKey should be printed as Bytes.toString(rowKey).

----=20

*object toString contain mi*

hbase-1.2.2/hbase-client/src/main/java/org/apache/hadoop/hbase/client/Async=
Process.java

{code}
LOG.warn("#" + id + ", the task was rejected by the pool. This is unexpecte=
d."+ " Server is "+ server.getServerName(),t);
{code}
server is an instance of class ServerName, we found ServerName.java:

hbase-client/src/main/java/org/apache/hadoop/hbase/ServerName.java
{code}
  @Override
  public String toString() {
    return getServerName();
  }
{code}
the toString method returns getServerName(), so the "server.getServerName()=
" should be replaced with "server" in case of simplicity and readability

Similar examples are in:

hbase-1.2.2/hbase-client/src/main/java/org/apache/hadoop/hbase/client/Preem=
ptiveFastFailInterceptor.java

{code}
LOG.info("Clearing out PFFE for server " + server.getServerName());
return getServerName();
{code}

hbase-1.2.2/hbase-server/src/main/java/org/apache/hadoop/hbase/ipc/RpcServe=
r.java

line 705:=20
{code} LOG.debug(getName() + ": disconnecting client " + c.getHostAddress()=
); {code}

line 1259:
{code}=20
public String toString() {
      return getHostAddress() + ":" + remotePort;
    }
{code}


> Several log refactoring/improvement suggestions
> -----------------------------------------------
>
>                 Key: HBASE-16469
>                 URL: https://issues.apache.org/jira/browse/HBASE-16469
>             Project: HBase
>          Issue Type: Improvement
>          Components: Operability
>    Affects Versions: 1.2.5
>            Reporter: Nemo Chen
>            Assignee: Nemo Chen
>              Labels: easyfix, easytest
>             Fix For: 2.0.0, 1.5.0
>
>         Attachments: HBASE-16469.master.001.patch
>
>
> *method invocation replaced by variable*
> hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/handler/C=
loseRegionHandler.java
> line 118: {code}String name =3D regionInfo.getRegionNameAsString();{code}
> line 142: {code}LOG.warn("Can't close region: was already closed during c=
lose(): " +
>             regionInfo.getRegionNameAsString()); {code}
> In the above two examples, the method invocations are assigned to the var=
iables before the logging code. These method invocations should be replaced=
 by variables in case of simplicity and readability
> ----
> *method invocation in return statement*
> hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegion.j=
ava
> line 5455:
> {code}
> public String toString() {
>     return getRegionInfo().getRegionNameAsString();
>   }
> {code}
> line 1260:
> {code}
> LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
>           + " is not mergeable because it is closing or closed");
> {code}
> line 1265:
> {code}
> LOG.debug("Region " + getRegionInfo().getRegionNameAsString()
>           + " is not mergeable because it has references");
> {code}
> line 1413:
> {code}=20
> LOG.info("Running close preflush of " + getRegionInfo().getRegionNameAsSt=
ring());
> {code}
> In these above examples, the "getRegionInfo().getRegionNameAsString())" i=
s the return statement of method "toString" in the same class. They should =
be replaced with =E2=80=9Cthis=E2=80=9D   in case of simplicity and readabi=
lity.
> ----
> *check the logged variable if it is null*
> hbase-it/src/test/java/org/apache/hadoop/hbase/HBaseClusterManager.java
> line 88:=20
> {code}
> if ((sshUserName !=3D null && sshUserName.length() > 0) ||
>         (sshOptions !=3D null && sshOptions.length() > 0)) {
>       LOG.info("Running with SSH user [" + sshUserName + "] and options [=
" + sshOptions + "]");
>     }
> {code}
> hbase-server/src/main/java/org/apache/hadoop/hbase/master/AssignmentManag=
er.java
> line 980:
> {code}
> if ((regionState =3D=3D null && latestState !=3D null)
>           || (regionState !=3D null && latestState =3D=3D null)
>           || (regionState !=3D null && latestState !=3D null
>             && latestState.getState() !=3D regionState.getState())) {
>         LOG.warn("Region state changed from " + regionState + " to "
>           + latestState + ", while acquiring lock");
>       }
> {code}
> In the above example, the logging variable could null at run time. It is =
a bad  practice to include null variables inside logs.
> ----
> *variable in byte printed directly*
> hbase-server/src/test/java/org/apache/hadoop/hbase/util/MultiThreadedUpda=
ter.java
> line 145:=20
> {code}
> byte[] rowKey =3D dataGenerator.getDeterministicUniqueKey(rowKeyBase);
> {code}
> line 184:
> {code}
> LOG.error("Failed to update the row with key =3D [" + rowKey
>       + "], since we could not get the original row");
> {code}
> rowKey should be printed as Bytes.toString(rowKey).
> ----=20
> *object toString contain mi*
> The toString method returns getServerName(), so the "server.getServerName=
()" should be replaced with "server" in case of simplicity and readability.
> hbase-client/src/main/java/org/apache/hadoop/hbase/client/PreemptiveFastF=
ailInterceptor.java
> {code}
> LOG.info("Clearing out PFFE for server " + server.getServerName());
> return getServerName();
> {code}


--
This message was sent by Atlassian JIRA
(v6.3.15#6346)