the *filename* and not the pid_file(!). Stupid brain-fault on my part.
This could cause a segfault under -neworder if newsyslog had to rotate
multiple files, and later ones had specifed the 'N' flag.
Bug first reported by: le
MFC after: 3 days
processes, and balance that by adding a 10-second delay after all the
processes have been signaled. Also improvement a few messages printed
with `-n' or `-v' processing (mostly signal-related messages).
MFC after: 13 days
files to rotate. The new order will first rotate all files that need
to be rotated, and then send a single signal to each process which
needs to be signaled, and finally it will compress all the files which
were rotated.
This means daemons will be signaled once per run of newsyslog, instead
of once per file rotated. Also, files will be compressed in order of
file-size (smallest to largest). Also, it waits for each file to be
completely compressed before starting the next one (effectively as if
the 'w' flag is specified for all entries in newsyslog.conf). This
avoids the situation of having 10 gzip's going at the same time (each
with a log.0 and a log.0.gz file active), and it also means that file
attributes can be reliably set on files after they are compressed.
NOTE: This commit does define NEWORDER (which you could get rid of if
you really don't trust this), but it does not flip the "-D neworder"
switch. So, at the moment none of these changes happen unless you
request them (perhaps by adding '<debug> neworder' in newsyslog.conf).
PR: bin/25070 inspired some parts of this
Submitted by: parts from bin/25070 done by Helge Oldach
MFC after: 14 days
the newsyslog.conf file. Rename one size-related variable, and move
another one from the stack into conf_entry. Add a routine to change
file-attributes (chown, chmod, chflags), instead of having several
places doing the same sequence of system-calls. A few cosmetic/style
changes.
These should not effect any users. Most of these probably look
pointless, but they are the "insignificant parts" of a much larger
update that I'll be committing soon. Doing these as a separate update
should make that update easier to read.
MFC after: 14 days
that had been written some months ago for other processing. This
should get rid of a few subtle situations where an existing log
file would not exist (for a short time) while it is being rotated.
MFC after: 16 days
"will trim at" message printed when the user requests '-v'. The
previous code would often print the wrong time, such as:
On Sept 22, run: newsyslog -nv /var/log/wtmp
And see: will trim at Mon Sep 1 05:00:00 2003
correct msg: will trim at Wed Oct 1 05:00:00 2003
MFC after: 20 days
would only match a leap year every 400 years. The parseDWM code first
showed up in April 2000, so the first time this bug would cause any
confusion is in Feb 2004.
MFC after: 18 days
only code-change is to add a "next_time" parameter to both routines (and
that is not used yet). A later update will make "next_time" more useful.
MFC after: 20 days
debugging options. Initial option is '-D TN=<time>', which can be
used to see how newsyslog would work if run at the specified time.
(time format is ISO 8601, since that is already supported).
MFC after: 23 days
was mistakenly calling the standard isnumber() function to find out if
the given 'user' or 'group' were all numeric. This meant that only the
first character of the fields were actually checked, so a username of
(say) '3com' would look like a number, and thus get mapped to uid=3 (bin)
instead of username=3com.
This bug was introduced back in freebsd's v1.1. That initial import
almost matches netbsd's v1.9, except that an internal isnumber()
routine was removed in favor of the standard library version. The thing
is, that internal routine was checking the entire string, and not just
the first digit. In OpenBSD, isnumber() was eventually renamed to
isnumberstr() to make the distinction more obvious, and I'm going to
follow that lead.
I believe this also happens to remove the last references to isnumber()
in the entire freebsd base system.
Obtained from: OpenBSD, by a long circuitous route
MFC after: 5 days
rotated and then compressed with bzip2 instead of gzip. Otherwise,
any file which had a time-interval specified for 'when' and also
specified the 'J' flag would be rotated every time newsyslog was run.
(this is a quick-fix, trying to beat the code-freeze for 5.1-release)
PR: bin/51519
MFC after: 1 week
config file. If the -C option is specified once, then newsyslog will create
any entries which specify the 'C' option. If -C is given twice, then
newsyslog will create all missing log files. Some of this code comes
from NetBSD, although this implementation does not exactly match theirs.
Reviewed by: freebsd-arch
MFC after: 10 days
the 'N' flag. These were coded in March as revisions 1.55 and 1.56
of newsyslog.c. I intend to MFC all the matching changes next week.
This also reorganizes the description of the 'flags' field to give
list of the valid flags, instead of a long paragraph explaining
each of the possible values.
Obtained from: NetBSD (in spirit at least, for -s and N)
MFC after: 1 week
a filename pattern, and also wrt filenames given on the command line.
Now if a file is listed as a specific entry, it will not *also* be
processed by an entry specifying a pattern. And filename-patterns
will now only match existing files (ignoring directories, etc).
MFC after: 3 weeks
will contain the pid for a process group. This means the file must
contain a negative value (as would be needed in the 'kill' commmand).
I still need to write man-page update before MFC-ing.
This started by rewriting the get_pid() routine. Later I looked at
what OpenBSD has, and included a few ideas from their send_signal()
routine. So, parts of this change are from OpenBSD, even though
OpenBSD does not actually have a 'U' flag.
PR: bin/28435
Reviewed by: no objections on freebsd-arch
MFC after: 3 weeks
warning message if -s is specified and it rotates a file that expects
to be compressed. This warning message is not printed if -R is also
specified, because we assume a -sR request is coming from the process
which would have been signaled, and that it has already released the
logfile.
Indirectly noticed by: sheldonh
and config-file entries which specify a filename-pattern (glob). It is
still not perfectly-right, but at least it isn't completely-wrong.
Reviewed by: no objections on freebsd-arch
MFC after: 3 weeks
MFC addendum: (or after the code-freeze of 4.x is lifted)
should rotate all files given on the command, even if they don't seem to
need to be rotated. This would be used by some other command that decides
the given log file(s) should be rotated, but wants the "how" of that rotation
to be determined by entries to newsyslog. Wes expects to change syslogd to
take advantage of this. Man page will be updated after we're sure this is
all working the way we want it to.
Reviewed by: no objections on freebsd-arch
MFC after: 3 weeks
MFC addendum: (or after the code-freeze of 4.x is lifted)
not send a signal to any processes. Also add a config-file flag of 'N' or
'n', which indicates that the given logfile has no process which needs a
signal when it is rotated. Both of these are based on changes NetBSD
has made, although the implementation is somewhat different.
PR: bin/36553 (2nd half)
Reviewed by: no objections on freebsd-arch
Obtained from: NetBSD (in spirit, at least)
MFC after: 3 weeks
given for -a did not exist, then newsyslog would always try to create
it, even if -n was specified.
2) When -a processing *does* create the directory, have it check the result
from mkdir(), and immediately error-out if that failed.
PR: bin/46974
MFC after: 3 weeks
specified at runtime, but that filename is not listed in the newsyslog.conf
file. This default-action can be changed by having a line in newsyslog.conf
with the filename of "<default>". Before this change, the program would
quietly ignore the given file. An update to the man page will be written
after I finish some other updates to newsyslog.c.
Reviewed by: no objections from freebsd-arch
MFC after: 3 weeks
`J' flag means that newsyslog should wait for previously started compression
jobs to complete before starting a new one for this entry. When it is used
along with the `G' flag, in the case when several logfiles match the pattern
and should be compressed, the newsyslog will compress logs one by one,
ensuring that only one compression job is running at a time.
This prevents newsyslog(8) from overloading system by starting several
compression jobs on big files simultaneously.
Sponsored by: Porta Software Ltd
MFC after: 2 weeks
the specified filename of the log to be rotated is in fact shell glob
pattern. In this case, all files matching this pattern will be rotated
using the same options. Useful in the case when there is no pre-defined
name for the logfiles (e.g. xtradius, samba etc).
Sponsored by: PortaOne Software Ltd
MFC after: 2 weeks
files are owned by the caller of newsyslog (usually root:wheel) even if
alternative ownerships were specified in newsyslog.conf.
Note that this is part of a wider problem which is fully addressed in
OpenBSD. Anyone with the time and inclination to incorporate the full
fix for the wider problem will receive no complaints from me and should
feel free to walk all over this delta.
PR: bin/36738
MFC after: 1 week
give an example of how to rotate logs at the beginning of the month.
Although they sound the same, since both of them rotate logs at the
beginning of the day, the former ended up taking place on, e.g., July
31 00:00 instead of the expected July 31 23:59. This is contraty to POLA.
Submitted by: Dan Langille <dan@langille.org>
configurable directory
- implement alternate and more flexible way to specify
logfile rotation time in addition to the ISO 8601
restricted format
- cleanup the source which was a mix of several styles
of persons who maintained it so far, ran through
knfom script got from bde.
Reviewed by: (in part) sheldonh and garyj
Joerg Wunsch suggested to do this to make the functional changes in
the next commit to newsyslog (which were run through the same script)
better visible.
compatibility. : is still the documented non-ambiguous approach. The
algorithm used will correctly parse david.obrien.staff as strrchar() is
used, and in my mind more people would use a ``.'' in the username than
the group name.
Convinced by argument and patch by: sheldonh (with slight changes by me)
