Write a simple little program to remove the entry from the lastseen DB. It is
pretty easy, couple of BerkDB operations.
>
>
>> But I have set lastseenexpireafter => "4320"; (3 days) in the common
>> control and there are still clients showing up that have been gone for
>> more than 3 days.
>>
Well, I added a command to do a "ps -ef " if the process count was out
of range, and it did do it. But it did not show more than one process
running:
!! Process count for 'XX' was out of promised range (2 found)
I: Comment: Verify the processes are running
Q: "...usr/bin/ps -ef ": root 2
Well, there are lots of things to think about. Unix is not just Linux, there
are many more out there and not all the authentication/user situations are the
same. You also have to think about Solaris and yellowpages, systems without
/etc/shadow, LDAP, DCE, Krb5, etc. The list is endless in the di
Good idea. I find the process overcounting more common. In my configuration, I
have a range for the process count and it reports when the process count is out
of range. I will add a "ps -ef" along with the report. It may not catch it as
it is not run exactly as the same time as the cf process co
Good to know. However, I get the miscounting when running cf-exec in the
background. It is not just process undercounting, I also see process
overcounting. It is very odd, I searched through the code and I can't see how
this can happen. I can't figure out how to make a testcase in which this wil
Greetings,
I think I hit upon the perfect circumstance to bust the hash table in
GetVariable (vars.c). I believe that the hash table is not going far enough
before reseting the increment as I end up in an infinite loop. See the comments
"//" below:
Debug("GetVariable(%s,%s): using scope '%s' f
Does using policy=>"overridable" help? I had a similar issue with duplicate
value with upgrading to cf4:
> Vars:
> 8
> 9 redhat::
>10 # List this variable first as the default.
>11 "services" slist => {
>12 "anacron",
>13
It is passed as a parameter.
bundle agent run_bundle(bundle_name) {
On Mar 12, 2010, at 8:57 AM, nwat...@symcor.com wrote:
> help-cfengine-boun...@cfengine.org wrote on 2010-03-12 08:54:28
>>
>> % cf-agent -K
>> Undeclared promise bundle "$(bundle_name)()" was referenced in a promise
>
> Is bu
Greetings:
This is invalid? I suppose it should be, it is a little nutty:
bundle agent run_bundle(bundle_name) {
methods:
"any" usebundle => "$(bundle_name)",
comment => "run bundle $(bundle_name)";
}
% cf-agent -K
Undeclared promise bundle "$(bundle_name)()"
Primarily on AIX, I have not noticed the issue on other OS'es (I do
not have Solaris in the mix, but that will be soon).
2010/3/11 :
> Matt what types of systems do you see this on? I've only seen it on a
> Sparc Solaris 10 host.
___
Help-cfengine ma
once and awhile it does
happen.
2010/3/11 :
> Matt Richards wrote on 2010-03-10 15:05:32:
>
>> I have seen this behavior as well. Not very common, but it happens
>> every now and then.
>
> Matt,
>
> Do you call cf-execd from a cron job? If so can you please provi
I have seen this behavior as well. Not very common, but it happens every now
and then.
On Mar 10, 2010, at 2:55 PM, nwat...@symcor.com wrote:
> Has anyone noticed extra processes being started in spite of promises?
> Currently my process list is
> ps -ef|grep cf-
>root 23518 1 0 13:46
Greetings,
I see that cf-3.0.4 is out there, but access if forbidden to download it. I may
have jumped the gun, but I am very excited.
I haven't been to the cfengine web page lately, but, Wow! Lots of work and new
stuff. Very impressive. Did ya happen to see Jarle Bjorgeengen's cfengine vs
pup
Mine is similar to how Neil does it. I use plain old CVS. Specifying tags on
the cf configuration to indicate which level (alpha, beta, prod, etc...) the
edit should be in:
cvs tag -F alpha failsafe.cf
I then can key off the tag to release different levels and release using cvs.
Something lik
This sound like the best safe method. A little more work to deal with the file,
but no race conditions or quirky platform specific items.
On Mar 4, 2010, at 11:00 AM, Christopher Browne wrote:
>
> I cope with cron edits via...
>
> - Pulling it out into a file (crontab -l)
>
> - Editing the fil
True, only if you have the full crontab available as a file. With cfengine, I
just edit the cron entry directly without pulling the whole file. But, that is
another way to do it. I suppose you could do a
crontab /var/spool/cron/crontabs/
Interesting, I never thought to do it that way. It might
I SIGKILL cron for AIX, it is set to respawn in /etc/inittab. Others I did a
different action (like Linux listed below):
bundle agent update_cron {
files:
"$(global.crontabs)/root"
comment => "Add cf-agent to root's crontab",
edit_line => add_cf_cron,
classes =>
; execresult("/usr/bin/cal $(month) $(year) |
awk '{print $7}' | grep -E "^[0-9]" | tail -1","useshell");
classes:
"Last_Saturday" expression => classify("Day$(last_day)");
On Wed, Mar 3, 2010 at 8:26 AM, Matt Richards wrote:
>
Interesting. I had something similar with how to schedule for the last
day of the month (ya, the shell commands are cheating a bit):
vars:
"year" string => execresult("/usr/bin/date +%Y","noshell");
"month"string => execresult("/usr/bin/date +%m","noshell");
"last_day" string
Greetings,
I was wondering if there was any reason to compile mySQL into the
community edition of cfengine. It seems like only the enterprise
edition uses it.
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo
:36 AM, Mark Burgess wrote:
>>
>> Matt, are you saying there is a memory leak, or that there is some weird
>> memory
>> resource allocation on AIX? Could you work around by restarting the server
>> once a week?
>>
>> Matt Richards wrote:
>>>
going fine.
On Feb 12, 2010, at 9:36 AM, Mark Burgess wrote:
>
> Matt, are you saying there is a memory leak, or that there is some weird
> memory
> resource allocation on AIX? Could you work around by restarting the server
> once a week?
>
> Matt Richards wrote:
>> G
Greetings,
cf-serverd core dumped on all my AIX servers after about a month of running.
After looking at the core dump, it seems it is a rather common issue.
cf-serverd over time seems to accumulate too much data to be held in a single
memory segment. Recompiling with the compiler flag "-bmaxd
Yikes, I don't have gdb.I compile cfengine using the native
compiler. I'll put it in the bug tracking system.
Thanks.
On Mon, Feb 1, 2010 at 10:46 AM, Mark Burgess wrote:
>
> Matt, can you use the bug tracking system and gdb please
>
> thanks
___
H
Greetings,
I am getting a core dump on a specific machine. It has been behaving
well on 3.0.3 for a month now:
Segmentation fault in PurgeLocalFiles at line 526 in file ""
could not read "files_interfaces.c"
(dbx) where
PurgeLocalFiles(filelist = (nil), localdir = (nil), attr = (...), pp =
(nil))
ooo, my interest is piqued!
On Jan 26, 2010, at 3:11 PM, Janet Bass wrote:
> Has anyone used Applescripts scripts in cfengine?
>
> I am not finding a lot of documentation about Cfengine and Mac OS X
>
> Thanks
> --
> Janet Bass
> Unix System Administration
> Manufacturing Engineering Laborator
I am looking at getfields and it seems that if you have a null field, the value
in the named array is not populated for that entry. For example:
bundle agent test {
vars:
"passwd_match" int => getfields("^root:.*","/tmp/testfile",":","userdata");
reports:
cfengine_3::
"
I agree, you can't learn other programming languages (like C) overnight.
On Dec 15, 2009, at 8:59 AM, Mark Burgess wrote:
>
> The syntax is not tough. It is incredibly simple. One pattern to follow for
> everything.
___
Help-cfengine mailing list
He
off track?
On Dec 11, 2009, at 8:03 AM, Matt Richards wrote:
> Well, it is still doing it. I think it is the cf-server actually causing the
> illegal cipher length in the client. I am using openSSL 0.9.8l. I don't have
> the client side dialog, as it is very random on which host
ere has discovered some problems with openSSL beta 1, and had
> to revert to
> 0.9.8 something. What version of SSL are you using?
>
> M
>
>
> Matt Richards wrote:
>> Well some good news and some bad news. cf-serverd did core dump last
>> night on th
Well some good news and some bad news. cf-serverd did core dump last night on
the policy host. I did get output from two clients at the same time.
one was from a regular cf-agent regular run (06:22:02 am):
Protocol transaction sent illegal cipher length
!! Authentication dialogue with x.xx.xx.
}
On Dec 9, 2009, at 2:30 PM, Mark Burgess wrote:
>
> Matt, could you try svn and see if this helps please.
>
> Matt Richards wrote:
>> I am a little closer now. I added some debugging information in
>> client_protocol.c (~line 338):
>>
>>
ark Burgess wrote:
>
> Perhaps you have access to some fancy tools, like purify, insight etc
> that might help debug this. It sounds like some kind of heap corruption.
>
> M
>
> Matt Richards wrote:
>> Well, I hate to say this, but I am still having this problem (svn 657
Oh, I dare. I dare.
On Mon, Dec 7, 2009 at 7:47 AM, Mark Burgess wrote:
>
> Cfengine patch level 666 -- download if you dare
>
> M
>
> --
> Mark Burgess
>
> -
> Professor of Network and System Administration
> Oslo University College, Norway
>
> Per
>
> Matt,
>
> I am glad to hear that this no longer troubling you, but I am unable to trace
> the source
> of a problem. If you see anything else, let me know.
>
> Matt Richards wrote:
>> Okay, after working on this one for a long time, I did noticed that
>> the
I did have those issues, until I upgraded flex and bison. flex-2.5.35 and
bison-2.4 work for me.
On Dec 3, 2009, at 3:15 PM, nwat...@symcor.com wrote:
> Does anyone have a working method for building cf3 on aix that they can
> share? For me configure seems for work but make step returns an err
Yep, that is how I have been doing them. I am just lazy.
On Dec 3, 2009, at 11:45 AM, Mark Burgess wrote:
>
> You can choose timestamp so that you can keep dated versions alive, and then
> tidy them
> after a certain time... That is less convenient maybe, but doable
___
Greetings,
By no means is this a priority in any way. Just a suggestion on an addition.
Add rotate to copy_backup so I could keep the last X versions of a copied file.
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailma
I should have looked at my traceback more carefully:
> ssleay_rand_bytes(buf = " \271`", num = 22), line 485 in "md_rand.c"
> RAND_bytes(buf = " \271`", num = 32), line 227 in "rand_lib.c"
num=32 and num=22? That shouldn't happen. Looking at the code it should be the
same as it is not modifie
I have tried 0.9.8j/k/l and 1.0 beta 4. All seem to do it, but only on the
policy host, not the clients. There is a mixture of 64bit and 32bit machines,
but from the SSL code, it doesn't appear like that would make a difference,
especially since I bind it to a 32bit only compile.
On Dec 2, 200
I didn't think cf-monitord was required, I am just taking stabs in the dark on
this one. cf-serverd has been crashing a few times a day now and it is driving
me crazy trying to figure it out. It has nothing to do with cfengine (as far as
I can tell), but with the low level SSL functions. Maybe
Greetings,
I am trying to track down what I think is an SSL bug and I am wondering if
cf-monitord is required to be running. I am not currently running it.
___
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/h
Greetings,
Something odd happened around in svn around 646-648. I seem to be having a DB
issue now:
!! Error scanning hashbase
cursor: DB_NOTFOUND: No matching key/data pair found
!! Error scanning hashbase
cursor: DB_NOTFOUND: No matching key/data pair found
^CReceived signal 2 (SIGINT) whil
43 matches
Mail list logo
