[Previous][Next][Contents][Index]

Generating and viewing reports

his chapter describes how to view reports and logs, how to set logging preferences, and how to send information to your email address.

Viewing the daily report

The server performs daily maintenance processes, which are reflected in the daily report. The daily report also includes information summaries of the server's performance and activities for the last day.

The daily report is generated at the same time the daily tasks are run. (You determine when the daily tasks are run. See Chapter 8, "Maintaining your Collabra Server.")

You can view the daily report at any time from within the Collabra Server interface.

  1. Choose Reports|View Daily Report.

  2. From the pull-down list, choose the daily report that you want to view.

  3. Click on View Report.

Note: You can also specify that the daily report be sent automatically to your email address. See "Specifying an address for mail notices" for more information.

The daily report shows you the following information. Note that some information is color coded (red, green, or yellow). For example, the file system capacity appears in green if you have a lot of disk space left; it appears in red if you are running out of disk space.

In addition, the daily report also shows you information about server statistics, including:

These statistics are also viewable by choosing Reports|View Server Statistics. See "Viewing and interpreting statistics" for more information.

Viewing and interpreting log files

You can view information about your Collabra Server in the following log files:

Each line in the log files represents one event. The log files show the date and time the message was sent, the name of the process that logged the entry, followed by the actual message in quotes.

In the server log, messages are numbered according to priority, with lower numbers representing higher-priority messages. (These numbers are not displayed in the server log form, however.) Highest-priority messages, usually indicating an error condition that requires some action on your part, appear in red.

In the discussion replication log, messages indicating an error condition also appear in red.

The following example, a typical message you might see in the server log, shows the message format:

3 [96/06/19 04:00.58] innd:ME new client starting: 
          "/news/ns-news-pb2/bin/news/bin/nnrpd -s -i 204.257.234.169"
The processes that frequently send messages to the log files are listed in Table 7.1:
Process

 Description

innd

 The main server process that handles all discussion group replications, listens to the specified news port, and responds to connections from newsreaders.

nnrpd

 The server process that spawns a thread for each newsreader it is handling. The nnrpd process is invoked by innd. (This document sometimes refers to nnrpd as "reader process" or "reader processes.")

nntpsend

 The process that batches discussion group replications for transmittal by the innxmit process.

innxmit

 The process that transmits articles to servers. This process is invoked by the nntpsend process.

indexer

 The process that indexes articles for full-text searching purposes.

newstime

 The process that schedules jobs, such as running news.daily, updating profiles, the nntpsend process, and so on.

ns-admin

 The administration server process.

ctlinnd

 A standalone program, which can communicate with innd, and can be used to perform some basic administrative functions outside of the Collabra Server interface.

The server log

The server log can contain information about most server activities, such as who connected to the server, what discussion groups were read, what articles were posted, and what problems occurred.

The information you see depends on the log level specified: Debug, Fatal, Error, or Notice. You can specify the log level by choosing Reports|Logging Preferences.

The server log is archived each time the server's daily maintenance routines are run.

To view the server log:

  1. Choose Reports|View Server Log.

  2. Open the pull-down list of archived logs and choose the log you want to view.

  3. Type the number of entries you want to view.

  4. If you want to find specific information, type a search string. For example, if you want to see information about which machines connected to your server, you can type "connect" in the search field.

  5. Click the View Log button.

The following sections provide examples of messages you might see in the server log and describe how to interpret these messages. Examples are provided for:

Client connection information
 Client connection information provides a record of who has been accessing your server. These messages are informational only and do not indicate an error or problem.

For example, the following message indicates that the main process (innd) is starting a new reader and handing off the connection to a reader process (nnrpd). The connecting client has the IP address 204.257.234.169:

3 [96/06/19 04:00.58] innd:ME new client starting:
          "/news/ns-news-pb2/bin/news/bin/nnrpd -s -i 204.257.234.169"
The next message indicates that the nnrpd process has accepted the handoff for the above connections (notice that the IP address is the same):

3 [96/06/19 04:00.59] nnrpd(5):Starting new reader, got socket: 
          22 and cmdline:
          "/news/ns-news-pb2/bin/news/bin/nnrpd -s -i 204.257.234.169" 
          from innd
The next message indicates that the host ts368.dialup.prem.net has successfully connected. (This host has IP address 204.257.234.169, so it is the same host as in the previous two messages.)

3 [96/06/19 04:01.02] nnrpd(5):ts369.dialup.prem.net connect
Normal client session and disconnects
 During a client session and when a client disconnects, the client process logs informational messages about the sessions. Some of this information is used by the daily report to generate statistics.

The following message indicates that the connected client read 3 articles in the discussion group netscape.navigator:

3 [96/06/19 04:01.02] nnrpd(2):245.dallas-3.tx.dial-access.att.net
          group netscape.navigator 3
The next message indicates that the client read a total of 3 articles in 1 discussion group during its session and then disconnected from the server:

3 [96/06/19 04:01.02] nnrpd(2):245.dallas-3.tx.dial-access.att.net
          exit articles 3 groups 1
The next message indicates that the client was connected a total of 1662.313 seconds, used 3.360 seconds of CPU time for user code, and 0.620 seconds for system code:

3 [96/06/19 04:01.02] nnrpd(2):245.dallas-3.tx.dial-access.att.net
          times user 3.360 system 0.620 elapsed 1662.313
Abnormal client disconnects
 Abnormal client disconnect messages indicate that a client is acting improperly. You can generally ignore these types of messages.

For example, the following message indicates that a client has attempted to connect to the news port, but was unable to do so:

2 [96/06/19 04:00.58] innd:ME cant accept RCreader Protocol error
You might see this message if a client that is not a newsreader attempts to connect to the news port. Because the server and the connecting program do not speak the same language (protocol), they cannot communicate, so innd closes the connection. In most cases, you can ignore this message.

The next message indicates the network connection has been closed due to a timeout:

3 [96/06/19 04:01.02] nnrpd(2):245.dallas-3.tx.dial-access.att.net
          timeout
A "proper" news client will send a "quit" message when it is done. However, some clients do not send quit messages. Normally, the server will keep the network connection open until it receives a quit message. However, if a certain amount of time has passed and the client has not sent anything to the server, the server closes the connection. This is called a timeout, which you can safely ignore.

Incoming discussion groups
Incoming discussion groups are articles sent to your server from other servers.

The following message indicates that the server news.foo.com connected and sent 31 articles to your server. Your server accepted 29, refused 1, and rejected 1.

[96/07/15 15:58.03] innd:news.foo.com:29 closed seconds 12 accepted
          29 refused 1 rejected 1
Outgoing discussion groups
 Outgoing discussion groups are articles sent by your server to other servers.

The following message indicates that your server is sending articles to another server (called secnews in this example). Your server offered 1 article, which was rejected:

[96/06/19 04:25.06] nntpsend:[28260:28261] innxmit -a -t180 
          secnews.netscape.com...
[96/06/19 04:25.07] innxmit:secnews.netscape.com stats offered 1
          accepted 0 refused 0 rejected 1
[96/06/19 04:25.07] innxmit:secnews.netscape.com times user 0.180
          system 0.060 elapsed 0.453
If only one article is rejected, you need not worry. If, however, all your articles are refused or rejected, you should contact the administrator of the other server. If the administrator of the other server indicates that they are not receiving your replicated discussion groups, check for the following status messages in your log file:

4 [96/06/19 04:05.04]
          innd:E:secnews.netscape.com:secnews.netscape.com.work 
[96/06/19 04:05.04] innd:secnews.netscape.com makebatch-->secnews.netscape.com.work
[96/06/19 04:05.04] innd:secnews.netscape.com closed
[96/06/19 04:05.04] innd:secnews.netscape.com opened secnews.netscape.com:23:file
These messages indicate that the innd process has created a batch file containing article information to send to the server secnews. Check to see if nntpsend and innxmit are running. If they are not, make sure secnews is configured as a replication host.

Server load information
The server process (innd) occasionally asks the reader processes (nnrpd) for status.

Each reader process provides information about the number of clients it is handling. Next, innd takes note of the active value for future load balancing efforts. The reader process messages are informational only and do not indicate an error or problem.

Server errors
Server errors generally indicate a problem with the server that must be investigated and rectified if necessary.

The following message indicates an error occurred when innd was attempting to write information to a sending host. The error number, 10054, indicates that the connection was reset, which usually means that the sending host closed the connection early:

2 [96/08/16 15:20.50] innd:257.0.180.15:7 can't write [Errno:10054]
If you see a lot of these messages and your incoming discussion groups from that site (257.0.180.15 in this case) seem to be lacking, contact the administrator of the other site and see why the connection is getting dropped.

The next message indicates that innd is unable to communicate with an nnrpd process that it expects is running:

2 [96/08/16 17:00.38] innd:unable to query nnrpd(0): 232
Either the nnrpd has died or is hung. Please report this problem to the discussion group, snews://secnews.netscape.com/netscape.server.news.

The following message indicates that your active file is bad. (The active file contains information about all the discussion groups your server carries.)

2 [96/06/29 04:15.11] innd:ME bad_format active
The active file might contain a formatting error, an invalid discussion group name, an illegal discussion group name, duplicate discussion group names, or incorrectly formatted lines.

The following message indicates that the two required discussion groups, control and junk, do not exist. Without these groups, the server will not run:

1 [96/06/24 09:29.43] innd:ME internal no control and/or junk group
You can add the groups by choosing Discussion Groups|Manage Discussion Groups or you can remake the active file by choosing Advanced Tasks|Recovery and clicking the Remake active file button.

The next message indicates that discussion groups listed in your active file have invalid names according to RFC1036, the Internet standards document that defines valid newsgroup names:

2 [96/07/09 11:14.37] ctlinnd:parseactive invalid newsgroup name
2 [96/06/21 20:23.29] ctlinnd:parseactive invalid newsgroup name
          "GPsemi.lincoln.test3": Illegal
The discussion group name might be in uppercase, for example, or might contain numbers, such as foo.com.752. For more information about valid discussion group names, see http://www.cis.ohio-state.edu/htbin/rfc/rfc1036 RFC1036.

You can rename the group or remove the group. Although it is not recommended, you can override groupname validity checking by changing the REQUIRE_VALID1036_NAMES parameter value in the nsnews.conf file from DO to DONT. If you do change the value, be aware that doing so might cause server problems.

The next message indicates that innd was unable to delete a file. Normally this is not a problem, but it may be a symptom of another problem. If you start to see a lot of these messages, check for other problems, such as problems with permission or ownership:

2 [96/07/07 21:10.34] innd:ME cant unlink
The next message indicates that innd was unable to link an article file. (Article files are linked when an article is crossposted to save space.) The linking can fail for various reasons, including one of the link names existing as a regular file. If you get the following message too often, try renumbering your active file (choose Advanced Tasks|Recovery)

2 [96/07/25 19:08.17] innd:ME cant link talk/politics/misc/965
The next message indicates a failure to write to the history file. The Collabra server stores information about all articles in a file called history. If the Collabra server cannot write to the history file, disk configuration problems and corruption of files is likely to occur. The Collabra server will be unable to expire articles or to accept new articles:

2 [96/07/25 19:08.18] innd:ME cant write history  [Errno:10009]
Check for permission problems and lack of disk space. After fixing the problem, you will need to restart your server. You might also want to rebuild your history files, but be aware that this can take several hours, depending on the size of the history file.

The next message indicates a failure to flush the log files. The server flushes the log files when running news.daily, for example, or when rotating log files. The server names the backup log file name.old:

inte2 [96/06/24 09:29.43] innd:ME cant fflush log [Errno:9]
The error code indicates that there is a bad descriptor. Try stopping the server and restarting it.

The discussion replication log

The discussion replication log provides information about incoming discussion groups, including information about what articles are accepted by your server, what articles are rejected by your server, and why articles are rejected. The discussion replication log is archived each time the server's daily maintenance routines are run.

To view the discussion replication log:

  1. Choose Reports|View Discussion Replication Log.

  2. Open the pull-down list of archived logs and choose the log you want to view.

  3. Choose the number of entries you want to view.

  4. If you want to find specific information, type a search string. For example, if you only want to view information about incoming discussion groups from a specific server, you can type the name of the server.

  5. Click the View Log button.

The following example indicates that on September 24, at the given time, your server was offered and accepted an article from the server news.foo.com. The plus sign indicates that your server accepted the article.

Sep 24 15:24.10.569 + news.foo.com <sad231dadas@news.moof.com>
          news.somecom.com
The next example indicates that on Dec 4, at the given time, your server was offered and rejected an article from the server stormy.foo.com. The minus sign indicates that your server rejected the article because stormy.dev.batman is an unwanted discussion group (the discussion group is not listed in the server's active file).

Dec 4 11:27:45.586 - stormy.foo.com <32A5CBF6.72E3@netscape.com> 437
          Unwanted newsgroup "stormy.dev.batman"

Viewing and interpreting statistics

 You can view information about the following:

To view server statistics:

  1. Choose Reports|View Server Statistics.

  2. Click one of the following options:

  3. Click one of the following options

  4. From the Limit lists pull-down menu, choose the number of entries you wish to view.

  5. Click OK.

Discussion group replication

 The process of exchanging and transferring articles between servers is called discussion group replication. You can view information about 1) articles sent to your server from other servers (incoming discussion groups) and 2) articles sent by your server to other servers (outgoing discussion groups).

Incoming discussion groups

For incoming discussion groups, you can view information about:

Outgoing discussion groups

 For outgoing discussion groups, you can view information about:

How to interpret the statistics

 You can use this information to determine if your discussion group replications are operating correctly. That is, you can determine if:

All articles posted to the server

 All articles posted to your server refers to all articles posted to your server by users and by discussion group replication.

Posts and crossposts

If an article is posted simultaneously to multiple discussion groups, only one copy of the article is stored in the spool.

For example, assume an article is posted to royal.pets.snakes, royal.pets.lizards, and royal.pets.turtles. One copy of the article is stored in the spool for royal.pets.snakes. The "copies" for the other discussion groups are links to the main article. Consequently, for purposes of statistics, the post to royal.pets.snakes is counted as one post whereas the "posts" to royal.pets.lizards and royal.pets.turtles are counted as crossposts. Crossposts are not considered when determining statistics about the size of articles received by your server.

Information you can view

You can view the following information about articles received by your server:

In addition, you can view the following information. For lists, the number of entries you see depends on the number you choose in the Limit lists box.

How to interpret the statistics

 You can use this information to determine:

User connections

 A user connection is a connection by a user (not a connection by a server). For user connections you can view:

You can use this information, for example, to determine:

You can also view information about user disconnections, including the reason for each disconnection. You can use this information, for example, to determine:

User posts

 A user post is a post to your server from a user (not a replication from another server). For user posts, you can view the following information:

You can use this information, for example, to determine:

User reads

 A read is when a user reads an article on your server. For reads, you can view the following information:

You can use this information, for example, to determine:

Viewing and interpreting disk usage

 To view information about disk usage:

  1. Choose Reports|Disk Usage and type the name of the directory.

    Note: Disk usage statistics for the default directory /spool automatically appear when you choose Reports|Disk Usage.

  2. Click Run Summary.

You can view information about:

By viewing the disk usage statistics, you can determine how many articles are contained in a discussion group hierarchy and how much disk space the hierarchy is consuming.

The statistics are shown for each directory in the spool. You will always see statistics for the following directories:

You can use this information to determine:

Specifying logging preferences

 You can specify the number of log cycles and the level of detail that you want in the log files. To specify your logging preferences:

  1. Choose Reports|Logging Preferences.

  2. In the Log Cycles field, type the number of log cycles. This is the number of archived copies of the daily reports and logs that the server will keep.

  3. From the Log Level pull-down menu, choose the level of detail that you want in your logs.

    Each log level provides increasing verbosity as follows: Fatal, Error, Notice, Debug.

  4. Click OK.

Specifying an address for mail notices

 You can specify whether you want to receive certain information by email. For example:

To specify where your notices should be sent:

  1. Choose Reports|Mail Notices.

  2. In the Admin email field, type the email address of the administrator.

  3. In the Mail hostname field, type the mail server hostname.

  4. By clicking the associated button, specify whether you want the following sent to your email address:

    You can always view the daily report from within the Collabra Server interface (Reports|View Daily Report), so you might not want to receive the report by email. It's probably a good idea, however, to receive emergency warnings by email, so you can take corrective action immediately.

  5. Click OK.

Note: To specify whether you want control articles emailed to you, choose Discussion Replication|Control Article Handling.

[Previous][Next][Contents][Index]

Copyright © 1997 Netscape Communications Corporation