You can try adding -fno-strict-aliasing to your CFLAGS.
- Original Message -
From: no-re...@cfengine.com
To: help-cfengine@cfengine.org
At: 5/01 7:18:43
Forum: CFEngine Help
Subject: Compile tokyocabinet AIX 7.1
Author: jonesy1234
Link to topic: https://cfengine.com/forum/read.php?3,25
Hello all,
Have built v3.2.3 on HP-UX B.11.31 using GCC 4.3.5 and have noticed that it
dumps core (SIGBUS). Recompiled with -g and it doesn't dump core anymore. Any
specific GCC compile/link flags to use here? I prefer to not use -g as the size
of the binaries/libraries are significantly large
Hi all,
Is there a way for cf-serverd to deny multiple connections from the same
client? Thanks.
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
Hello all,
I just noticed that on solaris/intel systems, there is no "solaris" built-in
hard/soft class. By definition, "solaris" is a unix operating system developed
by Sun Microsystems and is not specific to either the sparc or the intel
architecture.
Given cfengine already has other bu
Thanks everyone for your input and information. Yes, the flexibility of
cfengine allows the user to design whatever solution they require. I have
updated my configuration with an applicable solution and works as expected.
However, be that as it may, I would have thought that replace_patterns w
Thanks, I wasn't aware of the ordering within edit_line, so this information is
helpful. I guess this order is as good as any given the permutation of
delete_lines, insert_lines, replace_patterns could make different outcomes.
For some reason I felt that replace_patterns would be higher in the
Noticed that I get duplicate lines when cfengine is applying the promise of
replace_patterns and insert_lines.
For example, the line exists in a file but is incorrect. It appears that
replace_patterns corrects the entry, but insert_lines also inserts the entry as
well. Without having to go thr
A couple of issues seen:
sysinfo.c: line 807, variable "class" is not defined in scope
processes_select.c: lin 391, "MIN" is not defined, fails to link at link time.
Hope this info helps.
Jim
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
htt
Still no luck on getting replace_patterns such as:
replace_patterns:
"DUMPADM_SAVDIR=.*" replace_with =>
all("DUMPADM_SAVDIR=/var/crash/$(sys.host)");
"DUMPADM_CONTENT=.*" replace_with => all("DUMPADM_CONTENT=kernel");
"DUMPADM_ENABLE=.*" replace_with => all("DUMPADM_ENABLE=ye
Going once
Anyone else NOT having problems with replace_patterns and wildcard regex?
Or am the only lucky one?
- Original Message -
From: Jim Amanatidis
To: bug-cfeng...@cfengine.org
At: 2/08 13:46:16
Hi,
Having an issue with replace_patterns and I submitted case 470 on 02/08. Can
Thanks, any luck with the regex issue (case # 470)?
replace_patterns using any wildcarding (ie .*, [a-z]*) etc. all fail, and the
update is not made.
- Original Message -
From: no-re...@cfengine.com
To: help-cfengine@cfengine.org
At: 2/07 13:18:43
Forum: Cfengine Help
Subject: Re: CF_E
some more details or even make a bug report if you think
there is a problem.
On 02/05/2011 08:07 PM, JIM AMANATIDIS, BLOOMBERG/ 731 LEXIN wrote:
> I think its definitely 300 minutes, not seconds.
> I ran cf-agent (working on some edit_lines rules), I get this:
>
> community> Refere
hese.
On 02/05/2011 05:23 PM, JIM AMANATIDIS, BLOOMBERG/ 731 LEXIN wrote:
> Is there a way to override this in v3.1.4?
> ___
> Help-cfengine mailing list
> Help-cfengine@cfengine.org
> https://cfeng
Is there a way to override this in v3.1.4?
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
Anyway to prevent cfengine from writing anything to syslog?
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
Saw the other day that v3.1.4 may be due out soon? Just curious if there is an
ETA, thanks, Jim
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
I built 3.1.3.a build 1076, so far its gone all day without dumping core, looks
promising since it used to dump core almost hourly before. thanks
- Original Message -
From: Seva Gluschenko
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: HELP-CFENGINE@cfengine.org
At: 1/20 2:46:55
This
keep any 3.1.x version of cf-serverd running for longer than an
hour or so.
- Jim
- Original Message -
From: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
To: seva.glusche...@gmail.com
Cc: HELP-CFENGINE@cfengine.org
At: 1/20 6:46:24
Seva, thanks. Will pull down and build latest svn release.
--
Just built/installed v3.1.3, running cf-promises I get things like this:
!! Class expression "(any).!(bbh_dev)" has a missing operator in front of '('
!! Class expression "((Has_SR.bbh_routes)|(Has_HSR.bbh_routes))" has a missing
operator in front of '('
!! Class expression "(solaris).!(bbc_tr
Seva, thanks. Will pull down and build latest svn release.
- Original Message -
From: Seva Gluschenko
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: HELP-CFENGINE@cfengine.org
At: 1/20 2:46:55
This problem had been discussed recently in the list. The current svn
must have it fixed
Should have put this as well, (bad cut/paste):
...
Reading libavl.so.1
Reading libc_psr.so.1
t@156 (l@156) terminated by signal SEGV (no mapping at the fault address)
0xfe6b1d60: strcmp+0x0160: ld [%o1 + %o2], %o3
Current function is DeleteScope
Built 3.1.2 on Solaris 10 using Sun St
Built 3.1.2 on Solaris 10 using Sun Studio 12, keeps dumping core.
dbx output shows below
Anyone know of a fix for this issue? Thanks, Jim
(dbx) where
current thread: t@156
[1] strcmp(0x0, 0xff2b714c, 0xd48eb4, 0x0, 0x80808080,
fault. Only the cf_lastseen.db is now used.
Note the change in capitalization.
On 01/19/2011 05:02 PM, JIM AMANATIDIS, BLOOMBERG/ 731 LEXIN wrote:
>
> Hi,
> I've been using the same build process, so not sure why it may have stopped
> working when I upgraded to 3.0.5p1. But
Hi,
I've been using the same build process, so not sure why it may have stopped
working when I upgraded to 3.0.5p1. But to clarify,
/var/cfengine/cf_LastSeen.db updates every once in a while, but files such as
/var/cfengine/cf_LastSeen.db.HOSTNAME* do not update.
I am using the following re
I experienced the same issue running on Solaris 10, cfengine v3.0.5p1.
Any luck in finding out?
- Original Message -
From: Deb Heller
To: help-cfengine@cfengine.org
At: 1/18 21:35:25
I forgot to mention that I stopped cfengine processes, removed
cf_lastseen.tcdb - and restarted cfengi
Bidwell
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: HELP-CFENGINE@CFENGINE.ORG
At: 12/22 10:24:49
On Wed, 2010-12-22 at 09:31 -0500, JIM AMANATIDIS, BLOOMBERG/ 731 LEXIN
wrote:
> Having an issue with iprange not working in v3. Stripped out everything and
> reduced it to this code snippet
Note, ran this through with "--debug", seems that cf-agent does not pick up the
interface for the missing subnet?
# ifconfig aggr4
aggr4: flags=1001000843 mtu 9000
index 19
inet 10.125.7.98 netmask ff80 broadcast 10.125.7.127
ether 0:14:4f:de:4c:6
# grep "Checking IP" tm
Having an issue with iprange not working in v3. Stripped out everything and
reduced it to this code snippet (below), still not working. What gives?
# cat test.cf
body common control {
bundlesequence => { "test_routes" };
}
bundle agent test_routes {
classes:
"Has_Net1" expression => ipr
Hi folks,
Asked this previously, not sure if anyone has found a way to do it. But when
recursively copying directories that contain symlinks, on the 1st run the
symlinks may not get created because the target file hasn't been copied yet,
and the symlink is missing. True on subsequent copies t
self.
What is missing is the ability in "copy_from" to specify that symlinks (if
missing the source) should still be created.
Comments?
- Original Message -
From: Daniel V. Klein
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: HELP-CFENGINE@CFENGINE.ORG
At: 11/17 16:
Hi all,
Don't know why I haven't noticed this before, but when copying directories
recursively for the first time, and the source has symlinks in it, depending on
the sequence of how files are seen, the symlink does not get created if the
source file isn't copied first. This behavior is seen i
Hi folks,
I used to run v2.2.3, recently upgraded to v3.0.4, and received this error on
cf-serverd hosts:
cf-serverd[3602]: [ID 823470 daemon.error] !! Too many open databases
cf-serverd[3602]: [ID 823470 daemon.error] Fatal cfengine error: OpenDB: Could
not save DB handle
Checked source co
Anyone ever see this problem?
cf3 Will not create directories for a relative filename
($(sys.workdir)/lib.stage.v3/.). Has no invariant meaning
cf3 Can't make directories for $(sys.workdir)/lib.stage.v3/. in files.copyfrom
promise
files:
any::
"$(sys.workdir)/lib.stage.v3"
c
Hi folks,
Anyone see this issue before where cf-execd dumps core?
I opened a bug case on cfengine.org.
t...@null (l...@1) program terminated by signal SEGV (no mapping at the fault
address)
Current function is BDB_CloseDB
69 return (dbp->close(dbp,0) == 0);
Any help would be appreci
It would be a whole lot easier if returnszero just worked I think.
- Original Message -
From: Daniel V. Klein
To: becker...@mail.nih.gov
Cc: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN), help-cfengine@cfengine.org
At: 11/01 17:38:57
You could always try -s :-)
On Nov 1, 2010, at 5:21 PM
Okay, this seems to break with the string being searched is wrapped in quotes,
regardless of how many "\" surround it to escape it:
classes:
"Has_ce_nic" expression => returnszero("/bin/grep \\\"ce\\\"
/etc/path_to_inst >/dev/null 2>&1", "useshell");
Note, if I use "noshell" and remo
Hi, thanks for responding. The output creates noise. Under v2, output could be
redirected to /dev/null, does that functionality not exist in v3?
- Original Message -
From: Seva Gluschenko
To: marco.le...@domtar.com
Cc: HELP-CFENGINE@cfengine.org, JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN
Hi all,
In this example:
classes:
"Is_SOL10_U1"expression => returnszero("/bin/grep s10s_u1wos
/etc/release", "noshell");
Is there a way to redirect grep output to /dev/null?
Thanks, Jim
___
Help-cfengine mailing list
Help-cfengine@cfengin
Hello folks,
I submitted case #324 on cfengine.org on what I believe is a really bad bug
that surfaces in copy_from routines. My files syntax is:
"/bb/bin/syncfs"
copy_from => copyit("/SOURCE/BB/BIN/syncfs"),
perms => mog("500", "root", "other");
This winds up renaming /bb/bin to
Anyone know why this is being reported as error in syslog in version 3.x?
cf-execd[6136]: [ID 823470 daemon.error] Logical start time Wed Oct 20 23:33:57
2010
Under v2 it was daemon.notice (I use LOG_DAEMON as the facility). Stopping
cf-execd should be flagged as an error. Or is there any other
Hi all,
What do people use to stop cf-execd? I simply use kill PID, but every once in
a while cf-execd dumps core. Any suggestions would be appreciated. Thanks, Jim
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/
Is the correct way to reference HasCe inside config_etc_system this?
$(foo1.HasCe)
- Jim
- Original Message -
From: Jim Amanatidis
To: HELP-CFENGINE@CFENGINE.ORG
At: 10/05 10:25:41
Looking for some help/info on classes scope, for example:
bundle agent foo1 {
classes:
"HasCe"
Looking for some help/info on classes scope, for example:
bundle agent foo1 {
classes:
"HasCe" expression => returnszero("/bin/grep \\\"ce\\\" /etc/path_to_inst",
"noshell");
files:
"/etc/system"
edit_line => config_etc_system;
}
bundle edit_line config_etc_sys
Recently upgraded cfengine server to v3.0.4, I'm getting this in my syslog file:
cf-serverd[7753]: [ID 823470 user.error] Private decrypt failed = block type
is not 02
cf-serverd[7753]: [ID 823470 user.error] Private decrypt failed = padding
check failed
any idea folks? thanks, Jim
_
Hello all
Anyone have this issue, and is there a fix? Thanks - Jim
running v3, and notice it uses "/bin/ps -elf" to grap process table. However,
each line cuts off at 124 characters on solaris and it fails to match the
processes.
For example:
# /bin/ps -ef | grep cfengine_watch.sh
root 13
Never mind, found it as stealth. This v2=>v3 porting effort would go a lot
smoother if keywords were similar. timestamp=keep is intuitive. stealth=true,
is meaningless.
- Original Message -
From: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
To: HELP-CFENGINE@CFENGINE.ORG
At: 8/23 14:16
Hello folks
Is there an equivalent of funtionality in v2 of copy: "timestamp=keep" in v3?
I am setting owner and mode specfically, but want to preserve the timestamp
from the source file. Thanks - Jim
___
Help-cfengine mailing list
Help-cfengine@cf
Hello all,
Under v2, TrustKeysFrom had the ability to use networks, for example:
TrustKeysFrom = ( 192.168.0.0/16 )
this way not every IP had to be specified.
In v3, documentation doesn't show if this is still allowed or not, anyone
know?
Thanks,
Jim
___
Hello all,
Porting from v2 to v3, anyone know if there is an equivalent of
"IncrementPointer" when used as "IncrementPointer 3"?
Under location body, there is "select_line_matching". That part I got
working, but now I need to move the cursor down 3 lines to insert some text.
Thanks,
Jim
___
Hello all,
Is there a way to reduce timeout for selectservers function? Thanks, Jim
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine
ah, yes, thanks. sorry, brain meltdown looking at this too long.
- Original Message -
From: Wes Hardin
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: help-cfengine@cfengine.org
At: 7/28 11:13:43
move_obstructions doesn't belong in the link_from body, it's part of the paren
source => "$(x)";
move_obstructions => "true";
}
----- Original Message -
From: Wes Hardin
To: JIM AMANATIDIS (BLOOMBERG/ 731 LEXIN)
Cc: help-cfengine@cfengine.org
At: 7/28 11:01:09
Maybe you need "move_obstructions" then?
http://www.cfengine.org/m
Thanks, but that's not the issue I am having.
My source file already exists (it's a plain file)
My target file also exists.
With ->! I was able to nuke the plain file and establish a symbolic link to the
target file.
I looked at the reference documentation, the "when_no_source" parameter doesn't
Hello all,
Went through the documentation, I don't see an equivalent of forcing a
symbolic link to be created if the source file is not a symlink. For example:
/my_symlink =>! /target
Will ensure that /my_symlink is always a symlink to /target.
Is there a version 3 equivalent? I only found
Hello againg,
Thanks, but this doesn't work if I also specify any:: class as well.
For example:
nyprovdev1 # more promises.cf
body common control
{
any::
bundlesequence => { "do_generic", "do_specific_OS" };
inputs => { "generic.cf" };
linux::
inputs => { "linux.cf" };
Hello folks,
Was wondering what people do for reducing the amount of files that get read/
scanned in v3? In my v2 setup, my cfagent.conf file looked like:
import:
any::
cf.allsite
solaris::
cf.solaris
aix::
cf.aix
linux::
cf.linux
Is there an equivalent in v3? Looking to
56 matches
Mail list logo
