[Comp.Sci.Dept, Utrecht] Note from archiver<at>cs.uu.nl: This page is part of a big collection of Usenet postings, archived here for your convenience. For matters concerning the content of this page, please contact its author(s); use the source, if all else fails. For matters concerning the archive as a whole, please refer to the archive description or contact the archiver.

Subject: INN FAQ Part 5/9: Other error messages and what they mean

This article was archived around: 9 Dec 1997 03:25:40 +0100

All FAQs in Directory: usenet/software/inn-faq
All FAQs posted in: news.software.nntp, news.software.b
Source: Usenet Version


Posted-By: post_faq 2.10 Archive-name: usenet/software/inn-faq/part5 Last Changed: $Date: 1997/11/18 02:25:35 $ $Revision: 2.26 $
Part 5 of 9 INN FAQ Part 1: General and questions from people that don't (yet) run INN INN FAQ Part 2: Specific notes for specific operating systems INN FAQ Part 3: Reasons why INN isn't starting INN FAQ Part 4: The debugging tutorial (setup of feeds etc.) INN FAQ Part 5: Other error messages and what they mean INN FAQ Part 6: Day-to-day operation and changes to the system INN FAQ Part 7: Problems with INN already running INN FAQ Part 8: Appendix A: Norman's install guide INN FAQ Part 9: Appendix B: Configurations for certain systems ------------------------------ Subject: Table Of Contents for Part 5/9 ===================================================================== TABLE OF CONTENTS FOR PART 5/9 ===================================================================== OTHER ERROR MESSAGES AND WHAT THEY MEAN: 5.1 ld.so: Undefined symbol: _dbzwritethrough 5.2 Bad file number writing history file -- throttling 5.3 log file fills with "<site> <article-id> 436 No space" 5.4 Why does my innd often die with the message "Can't sync history..." 5.5 Reply from ctlinnd: Can't send XXX command (sendto failure) 5.6 syslog msg: ME cant sendto CCreader bytes 4 No such file or directory 5.7 syslog: cant symlink 5.8 syslog: cant read Connection reset by peer 5.9 inews says "bad message-id" 5.10 Why do all these "readclose" messages show up in my syslog? 5.11 "File exists writing symlinking article file -- throttling" 5.12 "cant fopen <newsgroup>/.thread No such file or directory" 5.13 news.daily reports: "Expire had problems removing articles" 5.14 syslog: nnrpd[22560]: ? cant gethostbyaddr 198.3.24.3 Permission denied 5.15 syslog: ME cant nonblock 15 Operation not supported 5.16 innd: ME cant renumber ... lo too wide 5.17 innd: ME cant update_active control 5.18 syslog message: innxmit[1234]: max connect failed Error 0 5.19 Can't open "/usr/local/news/shlock955", Permission denied 5.20 INND: PID file exists -- unclean shutdown! 5.21 cant fopen /usr/lib/news/distributions 5.22 cant fopen /usr/lib/news/distrib.pats 5.23 ME cant exec in /usr/lib/news/capture No such file or directory 5.24 ME cant remalloc xx bytes Cannot allocate memory 5.25 innxmit says: Ignoring line "news/group/name/art_num art_id" 5.26 innxmit says; Can't connect to news.foo.bar Permission denied 5.27 innd: s (s might be replace by other letters) (in syslog ) 5.28 innd: site: closed 5:closed seconds 12 accepted 3 rejected 4 refused 5 5.29 syslog: nnrpd[13]: '1.2.3.4' bad_auth 5.30 Not a directory writing article file -- throttling 5.31 syslog: nntpd[13]: cant open ...../subscriptions ... 5.32 syslog: nntpd[13]: post ok ^A 5.33 news.daily: Can't open a/b/.LCK.overview, No such file or directory 5.34 syslog: innd: site buffered 5.35 syslog: innd: cant bind RCreader Permission denied ===================================================================== OTHER ERROR MESSAGES AND WHAT THEY MEAN ===================================================================== ------------------------------ Subject: (5.1) ld.so: Undefined symbol: _dbzwritethrough > Everything compiles correctly, but when I try to test rc.local I get: > > hermes# sh /usr/local/etc/rc.news > ld.so: Undefined symbol: _dbzwritethrough > > What am I doing wrong? This means that you are using a $INN/lib/dbz.c file that hasn't been patched with the $INN/lib/dbz.pch patch. The patching is automatic. Running "make" should do the patch without you even noticing. To force the code to be re-patched (i.e. copy the original to dbz.c and apply the patch to it), simply delete $INN/lib/dbz.c and run the usual "make". The Makefile does everything for you. Why might the patch fail and leave you with an unpatched dbz.c? 1) You don't have Larry Wall's (or the FSF's) "patch" program installed. It is available at ftp://prep.ai.mit.edu/pub/gnu/patch-2.1.tar.gz 2) Most C++ compilers based on USL's cfront come with a program called "patch" that does something completely different than Larry Wall's (or the FSF's) "patch" program. If your path is configured so that this program is found before Larry Wall's (or the FSF's) "patch" program, you will get an unpatched dbz.c. To fix this, delete $INN/lib/dbz.c, change your path, and do the usual "make". If this file is deleted the Makefile will do everything for you. See also section "5.2 The DBZ package" of the Install.ms document. ------------------------------ Subject: (5.2) Bad file number writing history file -- throttling The system ulimit was set restrict files from growing beyond a certain size and the history file was to large for the limit. Check the ulimit setting for the news user. The ulimit should be set to infinite. You need to figure out how to set this for user "news" no matter if the user logged in, is running from cron, or was run at boot up. ------------------------------ Subject: (5.3) log file fills with "<site> <article-id> 436 No space" innd issues this message if it cannot create a directory for a group, _no matter what the reason for the failure_. In most cases, /usr/spool/news (and the mount point it is symlinked to, if any) accidentally has permissions and ownerships that do not permit write access by innd (which runs as "news"). After chowning them to news, and setting the permissions to 755, everything should work fine. ------------------------------ Subject: (5.4) Why does my innd often die with the message "Can't sync history, interrupted system call" Are you running SunOS? See "Known Problems" section of the installation manual." To the best of my knowledge, nobody has seen this problem on any other system. ------------------------------ Subject: (5.5) Reply from ctlinnd: Can't send XXX command (sendto failure) The socket sendto wants to write to isn't available. This can either come from a dead server (e.g. send a ctlinnd shutdown twice) or from a process (find -exec rm{} comes to mind) that removed the socket from the filesystem. Look after _PATH_NEWSCONTROL in config.data and find out if all components of the path exist and are accessible. (inncheck could be of help ...) This condition will occur whenever news.daily is run as root, leaving the ownership of history, active, etc files as root. Change ownership of these files back to news and then restart. ------------------------------ Subject: (5.6) syslog msg: ME cant sendto CCreader bytes 4 No such file or directory (Rich Salz replies:) It usually means that some ctlinnd command timed out and gave up before innd could get around to replying. Always a problem with datagrams. :-) Usually not a problem in real life however. In INN1.3, the timeout stuff is handled better so most of these should go away. You can ignore the messages, but if it bothers you, edit news.daily and find this line: ctlinnd -s -t`wc -l <${ACTIVE}` renumber '' 2>&1 You can rewrite it to be something like this: COUNT=`wc -l <${ACTIVE}` ctlinnd -s -t`expr ${COUNT} \* 5` renumber '' 2>&1 ------------------------------ Subject: (5.7) syslog: cant symlink This usually means an intermediate directory doesn't exist. Usually "ctlinnd renumber" clears this problem, but doesn't remove the cause. This is usually an indication of some larger problem. Often it means that you have opted to use MMAP on an operating system (like Ultrix) that doesn't have a working mmap() function. However, this would only be the cause if you ignored the Install.ms *and* part 2 of this FAQ. So, obviously, this isn't the cause :-). Another cause might be that "ctlinnd renumber" fails because of fields in the active being too narrow. See #5.17 for how to deal with this. ------------------------------ Subject: (5.8) syslog: cant read Connection reset by peer This means that the client vanished or deliberately aborted the nnrpd TCP stream rather than closing it gracefully. This is a common situation, and it does not normally indicate a problem. Most, or all, of these messages are due to clients crashing or being aborted in the middle of a dialogue with nnrpd. For example, a PC or Mac user shuts off their machine at 5pm. Or, you are seeing the results of a bad design decision in the NewsWatcher newsreading program for the Mac. Future versions of NewsWatcher will be (hopefully) more graceful about closing connections. In other words, they should send "QUIT\r\n" on the socket, then close it. ------------------------------ Subject: (5.9) inews says "bad message-id" If this is a Solaris 2.x system, you didn't delete the lines mentioned in "SVR4, Solaris 2.x, and SCO ODT 3.0" in part 2 of this FAQ. ------------------------------ Subject: (5.10) Why do all these "readclose" messages show up in my syslog? Chris Schmidt <cs@germany.eu.net> says: The "readclose" message indicates that a remote connection to your server was not correctly terminated with the server-command "quit". This can have two reasons. First the line your feed uses to connect to you might be instable so that the connection drops every now and then. Solution: either ignore theses messages or find out why the line is unstable. The second reason for these messages could be a missconfigured client-program at your feed. This means the program (e.g nntplink) does close the connection without sending the "quit" first. If you configure a lower number for the exit-timeout (-e) than the close-timeout (-C) in nntplink then exactly this will happen. Solution: ask your feed to fix its nntplink-setup. Let me repeat that: If you are using "nntplink" your -e value must be higher than your -C value. ------------------------------ Subject: (5.11) "File exists writing symlinking article file -- throttling" QUESTION: I'm running INN 1.4, and the server throttles itself, saying "File exists writing symlinking article file -- throttling". Why? I have no clue, other than to note that the message is being emitted while innd/art.c tries to link a crossposted group. ANSWER: Innd wrote the article to comp/foo/123 and then tried to symlink it to alt/bar/128 and found that the symlink failed with errno == EEXIST. This generally only happens when your active file does not match your file/directory use. The three most common cases of that are: Trying to use MMAP on Ultrix Trying to use MMAP on (older) Linux Some strange interaction with tind. If you are using Ultrix or Linux, turn off MMAP. You don't have a choice in this. The Ultrix mmap() function does something completely different than the Sun/BSD mmap() function. The Linux function gives you some of the functionality that Sun/BSD mmap() function has, but not enough. (The Linux people expect to have it fully up to spec eventually, yeah right.) At least one person has reported problems with ICL DRS6000 SVR4 Unix when using MMAP. Try turning off MMAP if you find problems. Some people have also seen this, when the filesystem has problems. Shutting down the system and running fsck helped in that case. It has been reported that tind writes to the active file and this confuses innd (innd assumes it is the only process writing to the active file). If you are using tin, upgrade to the newest version of tin, it can read the NOV database rather than the "tin" database. To fix the active file (which may be corrupted), make sure nobody else is writing to the active file, then do ctlinnd renumber '' to get things synchronized again. If your history file is corrupt, you should do: ctlinnd renumber '' makehistory -buv (or -bunv see #6.6 for it) ctlinnd renumber '' (Note: the "makehistory" will take hours to run. See #7.53 on this.) IF YOU IGNORE THIS PROBLEM LONG ENOUGH (by issuing the "go" command via ctlinnd) you will eventually get a "innd: ME cant update_active control" message in syslog. The cause of this error is dealt with elsewhere in this FAQ (#5.17). ------------------------------ Subject: (5.12) "cant fopen <newsgroup>/.thread No such file or directory" Q: nnrpd logs "cant fopen <newsgroup>/.thread No such file or directory" hundreds of times a day although I installed trn-3 and maintain an overview database. Why doesn't trn use overview files instead of mthreads data? A: trn-3 tries to open .thread files (or use XTHREAD) first because $spooldir/db.init still exists. Delete it. ------------------------------ Subject: (5.13) news.daily reports: "Expire had problems removing articles" This message tells you that you need to look in the file ${MOST_LOGS}/expire.log to find out what really happened. On the other hand... Expire reports this if it goes through the entire expire process and didn't find any articles to remove. It is normal to get this error the first few days you are running INN. For example, if the smallest integer that appears in the 4th field of your expire.ctl is a "4", then you're sure to see this error the first four days you get an news.daily report. However, those first four days are when you are still learning the system and it can be very shocking to see the error. "Eeek! Did I do something wrong?" Nope, news.daily is just telling you that you have a virginal system. Rich writes a more technical explanation: | When using the "delayrm" keyword, news.daily calls expirerm to | actually remove the articles that expire listed in its "-z" file. | As distributed, expirerm calls fastrm with the "-e" flag. This | flag says "exit non-zero if nothing was removed." In the normal | case, it is an error if expire doesn't find anything to remove. ...and a system being 4 days old isn't the normal case. So you get the error. Ignore it. ------------------------------ Subject: (5.14) syslog: nnrpd[22560]: ? cant gethostbyaddr 198.3.24.3 Permission denied When someone connects to nnrpd, nnrpd does a reverse lookup on the clients IP address so it can look up the hostname in nnrpd.access. If the client's DNS information is incorrect, you will get this error. Make sure your SOA record is correct, as well as your in-addr.arpa data. Derek <ddl@clipper.ssb.com> adds that another cause for this problem, at least under SunOS 4.1.x, is that nnrpd wasn't linked with the proper resolver library (statically or dynamically). Even when you've replaced Sun's fershlugger libresolv.a you've still gotta remind yourself to _use_ it properly. ------------------------------ Subject: (5.15) syslog: ME cant nonblock 15 Operation not supported I get the following "syslog" message in /var/adm/messages: Dec 2 20:40:04 venus innd: ME cant nonblock 15 Operation not supported Answer: (from paulr@umbc4.umbc.edu (Paul Riddle)) It turns out that this is happening because /usr/spool/news on the machine running innd is an NFS-mounted filesystem, and innd is trying to do an FIONBIO on my feed file, which is under /usr/spool/news/out.going. (tal@plts.org adds:) All news transports (INN, C news, B news) want the spool partition to be local. Newsreader can read from an NFS mounted partition without any problems but innd should only see local partitions. NFS has a blatant disregard for many of the file semantics that are needed for a good netnews implementation. If you don't agree, please feel free to prove the authors of B news, C news, and INN wrong. Include source code. :-) Greg Andrews <gerg@netcom.com> adds: The spool can certainly be kept on an NFS filesystem without anything breaking, as the FAQ suggests. I've kept the spool on an NFS server, split between two NFS servers, and split between a local filesystem and two NFS servers. Reads and writes over NFS tend to be slower than over local disk, even if the NFS backbone is 100BaseT or CDDI. The out.going filesystem can be put on an NFS server, but the file locking becomes a problem on some Unixes. I bypassed that by borrowing the lockfile technique I saw in the old "cleanup" release of C-News, and writing a replacement for nntpsend in Perl. I also eliminated the numerous 'ctlinnd flush' commands that can choke a busy server. The lockfile technique is pretty simple and seems to work well for multiple machines mounting a common out.going filesystem via NFS. The nntpsend replacement script creates a lockfile unique to itself (using the machine name and PID), and locks sites by hard linking site lockfiles to its own lockfile. The newslib filesystem can also be put on the NFS server, though the only advantage is a minor one: If everything's on the same NFS server, you only need to mount one filesystem on the client machines. Systems without unix-domain sockets sometimes see this error. Just ignore it. ------------------------------ Subject: (5.16) innd: ME cant renumber ... lo too wide When you try to renumber your active there are articles in your spool with numbers higher than what will fit in the fields in the active. Look under innd: ME cant update_active control below for how to fix it. ------------------------------ Subject: (5.17) innd: ME cant update_active control What does "innd: ME cant update_active control" mean? Look at your active file. One of the fields is "99999" and has to be incremented to "100000" but there is no space. Shut down innd ("ctlinnd shutdown x"). Edit your active file (see the "Safe way to edit the "active" file?" question in INN FAQ part 6) to add more leading zeroes to all the numbers. You can use the following for this: 1. ctlinnd throttle 'updating active' 2. awk '{printf "%s %.10d %.10d %s\n", $1, $2, $3, $4}' active > active.new if your awk printf spaves instead of leading zeroes in this step, then try the following instead of step 2: awk '{printf "%s %010d %010d %s\n", $1, $2, $3, $4}' active > active.new or perl -lane 'split;printf "%s %.10d %.10d %s\n", @F' active > active.new 3. mv active.new active 4. ctlinnd reload active updated 5. ctlinnd go 'updating active' See also #5.11. ------------------------------ Subject: (5.18) syslog message: innxmit[1234]: max connect failed Error 0 John Line <jml4@cus.cam.ac.uk> writes: If you get syslog messages like "innxmit[1234]: max connect failed Error 0" when using nntpsend, it probably means you messed up a line in nntpsend.ctl (specifically, missed out one of the first two fields). While nntpsend.ctl is an obvious place to look for an nntpsend problem, there is nothing obvious to link the error message directly to the problem, because the text "max" is actually something invented by nntpsend when processing the file, and doesn't exist in nntpsend.ctl. It means the next-to-last field was null, but was the second field when it should have been the third! NB Remember to try inncheck when you have problems like this. I only just thought of it, after finding the problem the hard way, and it immediately reported "nntpsend.ctl:18: malformed line." ------------------------------ Subject: (5.19) Can't open "/usr/local/news/shlock955", Permission denied This usually means you don't have /usr/local/news owned by "news". The first time you run "make install" it should set the proper ownership if you run "make install" as "root". ------------------------------ Subject: (5.20) INND: PID file exists -- unclean shutdown! This just means your machine crashed or you didn't shutdown innd properly and THEREFORE rc.news is doing a bit of cleanup work before it starts innd. There is not much for you to do. You might run makehistoy or fixhist.pl to find articles that are on disk, but not in history. ------------------------------ Subject: (5.21) cant fopen /usr/lib/news/distributions Step 1: Can you "cat" this file? Step 2: You can't? Does it exist? Step 3: It doesn't? That's because you have to create it yourself. It should list every distribution your users might want to know a description of, starting with the local one. Step 4: Here's a sample file to start from. It's appropriate for a New Jersey site: nj New Jersey ny New York nyc New York City ne North East, USA usa United States of America na North America ------------------------------ Subject: (5.22) cant fopen /usr/lib/news/distrib.pats A sample distrib.pats comes with the distribution and should be in your $inn/site directory automatically. The sample needs the following modification: Remove the line: 10:news.software.nntp:inet News.software.nntp has moved from inet groups to normal groups end 1994/begin 1995. If you get this error it sounds like something was badly configured on your machine. ------------------------------ Subject: (5.23) ME cant exec in /usr/lib/news/capture No such file or directory The newsfeeds file includes lines that looks like this: ## Capture all Foo, Incorporated, postings #capture\ # :*/foo\ # :Tp,H2:/usr/lib/news/capture %s That's just a sample of how one might capture articles. There is no capture program that comes with INN. It's just an example of how to use the Tp,H2 flags. Don't uncomment anything that you don't want to activate! ------------------------------ Subject: (5.24) ME cant remalloc xx bytes Cannot allocate memory This probably means that innd wants to grow bigger than what the actual limit for a single process is. To solve this either rebuild your kernel with a different default max, or add "unlimit datasize" to the beginning on your rc.news file. On some OS (e.g. FreeBSD 2.1.x) there might be some problems when using the supplied malloc. Linking with -lgnumalloc helped (after Joe Greco <jgreco@ns.sol.net>). For HP-UP see also 'HP-UX tips' in part2 of the FAQ (#2.6). ------------------------------ Subject: (5.25) innxmit says: Ignoring line "news/group/name/art_num art_id" Q: Innxmit logs the above error message - what does this mean? You set the parameters for the feed wrong in newsfeeds - try Wnm instead of Wmn. See also newsfeeds(5) and innxmit(8) for further information. ------------------------------ Subject: (5.26) innxmit says; Can't connect to news.foo.bar Permission denied If you get this error, then first look if permissions of all files are correct. If yes try out innxmit <ip-number of news.foo.bar> foo.togo If this works you have some problems with name resolution. You might use the ip-number in nntpsend.ctl instead of the hostname next time. ------------------------------ Subject: (5.27) innd: s (s might be replace by other letters) (in syslog ) Q: What do these innd: s (or other innd: <letter>) mean? Innd logs to syslog when it receives a command from ctlinnd. The letter corresponds to the command that is send to innd. For a list of commands, refer to include/inndcomm.h ------------------------------ Subject: (5.28) innd: site: closed 5:closed seconds 12 accepted 3 rejected 4 refused 5 This isn't really an error but an informative message of inn. It also appears with 'checkpoint' instead of 'closed'. This message tells you how long a connections from the remote existed, how many articles were accepted from this remote site, how many innd already had of those offered and how many were rejected of some reason (see below). If you get a lot of rejects from a site then you should investigate, as there might be some problems. Erland Sommarskog, <sommar@algonet.se> explains further: Rejected articles are articles in groups that you don't have in your active file, or you ban in hosts.nntp. This could be groups you have missed to add, it could be groups you don't want. Also articles which are already too old for you to accept (-c switch to innd) fall in this part. If you get many rejects then you might also check out the clocks on the sender and receiver machines as it might be that with wrong clock settings articles get refused which seem to be to old or posted in future. Refused articles are articles that are already present in your history file, and which for some reason are being offered to you again. If you have more than one incoming feed, you are likely to get a fairly high number in the Refused column. If you are a leaf site with only one incoming feed you can still get high numbers in this column, if you have a longer expiration time (including /remember/) than you feed has. ------------------------------ Subject: (5.29) syslog: nnrpd[13]: '1.2.3.4' bad_auth A Newsreader that could be allowed via nnrp.access did send bad authinfo commands. This could either be that the client has a password in nnrp.access and sends a wrong one or the client needs no password and sends one even if it does not need to. ------------------------------ Subject: (5.30) Not a directory writing article file -- throttling This typically means that in the spool there is a directory in the spool that thinks it is a file. This seems to happen when the machine running innd crashes, and the spool gets corrupted. Fsck then tries to make sense of the mess it finds; sometimes it guesses wrong. Here is a little script that finds the offending file and disposes of if: #!/bin/sh # # change this line to your news spool NEWS_LOC=/var/spool/news find ${NEWS_LOC} -type f -print | # list all files tr -d [0-9] | # get rid of numbers grep -v "overview" | # get rid of .overview files egrep -v "/$" | # everything w/ / at end of line sed -s "s/^/rm /" | # put rm in front of file sh -s # it's history or alternatively: find ${NEWS_LOC} -type f -name "[a-z]*" -print | xargs rm (From: Carlso Castro <carlos@mci.net> and Matthias Urlichs <urlichs@noris.net>) ------------------------------ Subject: (5.31) syslog: nntpd[13]: cant open ...../subscriptions ... This is a new option in unoff4. First look if _PATH_NNRPSUBS is correctly set. If yes, then just fill it with information ... . The subscriptions file contains what the NNTP "LIST SUBSCRIPTIONS" command returns. Some newsreaders, such as tin, use this command to get a default set of groups that people are subscribed to if they do not have a .newsrc. If you don't want to use it, there are no side-effects that I know of beyond the messages in your log. ------------------------------ Subject: (5.32) syslog: nntpd[13]: post ok ^A Q: What is this? Normally posting lines look like nntpd[13]: post ok <id>. A: Your posting was posted to a moderated group and is now sent by mail to the moderator who will (or not) post it to news. See also: #6.7, #6.8 ------------------------------ Subject: (5.33) news.daily: Can't open a/b/.LCK.overview, No such file or directory Expireover goes through the active file and tries to remove old entries of overchan records from the spool. If now a group exists, that is new i.e. that exists in active, but for which no articles have been received, then there is no spool dir for that group. Expireover tries to make a lock file in the directory, which fails, so it prints the above message. See also #4.19, #6.30, #7.21 ------------------------------ Subject: (5.34) syslog: innd: site buffered I run INN 1.5 and see the following in syslog - what does it mean? | Jan 9 03:10:02 cactus innd: news.foo.bar.com buffered It means either your newsfeeds file has too many entries to handle in the normal way, or you've run out of file descriptors. This isn't fatal, but a warning message. ------------------------------ Subject: (5.35) syslog: innd: cant bind RCreader Permission denied This means that innd can't bind to port 119 as either inndstart is not suid root or not started as root. Normally the install should take care of this. -- See <a href="http://www.netbsd.org">NetBSD</a> for a multiplatform OS What would you call a BBS run by a mom? A "mother board".