Linux Network Administrator Manual (19)
2000-07-31 15:49
Publisher: NetBull Readings: 971 Translation: Zhao Wei GoHigh@shtdu.edu.cn Chapter 19 NewsReader's Configuration News Reader (NEWSReader) is a method for providing users with convenient access to news system functions, such as mailing The article gets a list of newsgroups in an appropriate manner. The quality of this interface is facing endless war. There have been many news readers to transplant them on Linux. Below, I will narrate the basic setting method of three most common news readers named TIN, TRN and NN. One of the most effective news readers is $ find / var / spool / news -name '[0-9] *' -exec cat {} /; | More this is un * x reads their news stubborn method. However, most news readers are more complicated. They usually provide a full screen interface and display all groups of users subscribed by different levels and display a profile of all articles in a group and display a single article. At the level of the newsgroup, many news readers list the list of articles, showing the title of these articles and the author. In a large group, the user wants to understand the relationship between articles, although it is confirmed that responding to early articles is coming. A response article usually repeats the topic of the original article, just before the topic, "RE:". In addition, the Message ID of the following article may be given in References: Title. Sort by these two standards can be generated (actually a tree structure), which is often referred to as a clue (Threads). One of the tasks to write news reader programs is to design a scheme for generating clues because the time it takes is proportional to the square of the number of articles. Here, we will no longer discuss how the user interface is established. All news readers under Linux have a good help function, so you should explore it. Below, we will only involve management tasks. Most of them are related to the establishment of the clue database and the account number. 19.1 TiN Configuration The most universal news reader for clues operation is TiN. It is model IAIN LEA and is a very old news reader TASS. [1] When the user enters a newsgroup, it makes the clue operation unless you make this operation through NNTP, otherwise this clue operation is very fast. On a 486DX50 machine, when reading the article directly from the hard disk, it takes about 30 seconds to index 1000 articles. This operation takes about 5 minutes to the time of 5 minutes to pass NNTP to a news server. [2] You can update your index files at regularly by using the -u option or call TIN using the -U option to improve this operation. Typically, TiN is placed in the user's home directory under. Tin / INDEX. However, this may be expense, so you will want to save a copy of them at a central location. For example, this can be done by letting Tin SetUID for NEWS or a completely free account. [3] At this point, TIN will put all the clue databases under /var/spool/news/.index. For any file access and shell exit, it will reset his valid UID to the true UID of its user. [4] A better solution is to install the TIND index background program, which will run as a background program and regularly update the index file. However, this background program is not included in any Linux release, so you have to compile it yourself. If you are running a local area network with a central news server, you can even run Tind on the server and retrieve this index file by all customers through NNTP. Of course, this needs to be extended to NNTP. A patch of NNTPDs that implements this extended NNTPD is included in the TIN source program.
The TIN version contained in some Linux distribution is not compiled into NNTP support, but most TIN versions are now. When as a RTI is called or using the -R option, TIN will try to connect to the NNTP server specified in the file / etc / nntpserver or by the NNTPServer environment variable. The NNTPServer file is just simply in a row with the name of the server. 19.2 TRN Configuration TRN is also a header of an older news reader called RN (meaning reading news). The "T" represents "Lead Operation" in this name. It is written by Wayne Davidson. Not like TiN, TRN does not provide the generating function of the clue database of the runtime. Instead, it uses a clue database prepared by a program called MTHReads. This program must periodically update the index file from the cron. However, do not run MTHREADS does not mean that you can't access a new article, it just means that you will have all of the "Novell Buys Out Linux !!" articles in your article selection menu, not a single one You can easily skip the clues. In order to turn on the clue operation for a specific newsgroup, you must use a news group list to call mthreads on the command line. This list list is exactly the same as the composition of the SYS file: mthreads comp, rec,! Rec.games.go This command will turn on the clue operation of all groups in CoMP and REC, except for Rec.Games.go (play Go people do not need popular clues). Thereafter, you only need to call it online to operate from any options. Cancel operations for all groups in the Active file can be opened by calling MThreads using the ALL group list. If you receive news in the evening, you can usually run MTHREADS in the morning, but you can also run multiple times according to the required. Sites with great transport traffic may want to run MTHREADS in a later way. When it is running with the -d option during the system startup, it puts itself in the background, wake up every 10 minutes to check if there is a new to the article, if there is a clue operation. In order to run MTHREADS in the ensemble, write the following lines in your rc.news script: / usr / local / bin / rn / mthreads -deav -a option allows MTHREADS to automatically turn on the clue operation of the new group; -V is used The log files for MTHReads MT.Log use detailed log messages. The log file is located in the directory you install TRN. The old files that no longer exist must be deleted from the index files periodically. By default, only numbers will be deleted in the article below the low water tag. [5] But no matter how you have expired in this number or more (because the oldest article has given a long expired date in the Expires: Title field) can force running by giving MTHREADS a -e option. "Strengthened" expired operation to delete. When MTHReads run in an inner manner, the -e option makes it a reinforcing expiration operation at midnight in midnight. 19.3 NN's configuration by Kim F. Storm is a final goal is not to read news readers. Its name represents "no news", its motto is "No news is good news, NN is better." ("No news is good news. Nn is better.") In order to achieve this ambitious goal, NN belt There is a large set of maintenance tools that can not only generate clues, but also check, account management, use of statistical parameters, and access restrictions. There is also a management program called NNAdmin that allows you to implement these tasks interactively. It is very intuitive, so we will no longer discuss these aspects, but only involve the generation of index files. Nn's clue database management program is called NNMASTER.
It is usually started from rc.news or rc.inet2 scripts and runs as a background program. It is the clue operation on all groups in the Active file in the following way. Similarly, you can also periodically call nnmaster from the cron to give it a list of operations. This list is very similar to the subscription in the SYS file, but it uses spaces to replace a comma. It uses an empty parameter "" to represent all groups instead of using a pseudo group name ALL. An example of a call is as follows # / usr / local / lib / nn / nnmaster! Rec.games.go REC Comp COMP Note that the order here is important: the leftmost matching group is preferred. So if we will! Rec.games.go placed behind the REC, then all articles in this group will be operated by the clue. Nn provides several ways to delete expired articles from its database. The first method is to update the database by scanning the newsgroup directory and discarding an entry that does not exist in the corresponding article. This is the default operation of calling the NNMASTER using the -e option. Unless you are done by nntp, it is very fast. The second method is very similar to the default expiration operation of MTHReads, which only deletes those numbers below the Active file. It can be turned on with the -e option. Finally, the third method is to give up the entire database and re-collect all articles. This can be done by giving -e3 to nNMaster. The list of an expired group is given by the -f option in the same manner above. However, if you already have NMASTER run as a background, you must kill it (using -k) before the expiration start start, and then re-run it with the original option. This uses the first method to run an appropriate order that runs expired in all groups is: # nnmaster -kf "# nnmaster -lrc] There are still a lot of logo to fine-tune NN behavior. If you involve deletion of bad articles or classify the article, please read the NNMASTER's manual. NMASTER relies on a file called Groups, which is located in / usr / local / lib / nn. If it does not exist, it will be created. For each news group, it contains a line starting with the group name, the group name follows one optional time tag, and some flags. You can edit these logo items for the corresponding group to enable certain performance, but you'd better do not change the order of each group. [6] The licensed flag and the role of these logos are also described in detail in the NNMASTER manual. Note [1] Written by Rich Skrenta. [2] If the NNTP server itself is operated by the clues, and let customers retrieve the clues database, things will have a huge progress; for example, Inn-1.4 is doing this. [3] However, do not use Nobody for this. As a rule, no file or command is associated with this user. [4] This is why you get an error message when you call it as a super user. But no matter what, don't do it with root. [5] Note that C News does not automatically update this low water tag; you must run updatemin to update. See Chapter 17. [6] This is because their order must be consistent with the order of the entries in the (binary) Master file. source:
Linux free pigeon