KS-Soft. Network Management Solutions
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister    ProfileProfile    Log inLog in 

Check all HD volumes for free space (variable threshold)
Goto page Previous  1, 2, 3, 4  Next
 
Post new topic   Reply to topic    KS-Soft Forum Index -> Library
View previous topic :: View next topic  
Author Message
greyhat64



Joined: 14 Mar 2008
Posts: 246
Location: USA

PostPosted: Mon Mar 02, 2009 4:54 pm    Post subject: Reply with quote

Alex,
Good for you! I didn't realize you guys were 'listening in'.
As you can tell, it's nothing more than an elaborate WMI query.

Paul,
Can you elaborate on what led you to institute the "non-simultaneous execution" option?

Regards,
Joel
Back to top
View user's profile Send private message Send e-mail
KS-Soft



Joined: 03 Apr 2002
Posts: 12790
Location: USA

PostPosted: Mon Mar 02, 2009 5:10 pm    Post subject: Reply with quote

BTW: HostMonitor always use "non-simultaneously test execution" option for Active Script (not Shell Script) test method due to some problem with Microsoft scripting engine.

Regards
Alex
Back to top
View user's profile Send private message Visit poster's website
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Tue Mar 03, 2009 3:13 am    Post subject: Reply with quote

Quote:
1. I like what you've done, but what did you think of my idea to 'bury' the passing of 'Bad Argument' results in the ParseArg function?
Ah! I now see where you were going with that comment. I hadn't thought too much about it because the bad parameter test was a single line in the if statement, until I added the extra test on line 84. I will re-visit that one.

Quote:
2. Good find on the "shell script" handles issue. I think I read somewhere that the actual limit is 6, but it probably starts choking before then.
I found this because we have 60 servers to check and we are still running the eval license - the monitor stops periodically and when you re-start it all the disk test run at once. This led to about 30 servers timing out and reporting errors, so I changed the notification to run after 2 bad results and changed the foldre properties - we already had the disk tests in a site folder.

I will check out Active Script and see if the test can run there, rather than Shell Script.
[Edit] The reason I didn't use Active Script is you cannot pass parameters to the test. Shell Script has a nice little "parameters" box. Maybe a feature request?

cheers, Paul
Back to top
View user's profile Send private message
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Tue Mar 03, 2009 9:32 am    Post subject: Reply with quote

Joel,
I have posted the code on page 1 to save multiple pages of code.

The ParseArg function now does all of the checking and reporting, although it does make it a little more cumbersome having to drop through multiple if statements.

KS-Soft, I noticed that the "correct" response to enable Log Analyzer to work is "value space units", e.g. "10 GB". Does the %laststatus% - %status% test take this into account?

cheers, Paul
Back to top
View user's profile Send private message
KS-Soft



Joined: 03 Apr 2002
Posts: 12790
Location: USA

PostPosted: Tue Mar 03, 2009 7:23 pm    Post subject: Reply with quote

Quote:
KS-Soft, I noticed that the "correct" response to enable Log Analyzer to work is "value space units", e.g. "10 GB". Does the %laststatus% - %status% test take this into account?

Sorry, I do not understand the question. %Status% and %LastStatus% variables represent status of the test, not reply value. Status cannot be "10 GB", it can be "Ok", "Bad", "Host is alive", "No answer".

If you are asking about %Reply% variable (not %Status%) that you want to use in some expressions, then make sure you are using quotation marks for the variable ("10 Gb" will be converted to the number, 10 Gb will be not)

Regards
Alex
Back to top
View user's profile Send private message Visit poster's website
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Wed Mar 04, 2009 3:51 am    Post subject: Reply with quote

Sorry, my bad.
I can see now I can use %Reply% or %Reply_Number%.
I will modify the script to always return the value type, in line with the Log Analyzer requirements.

cheers, Paul
Back to top
View user's profile Send private message
greyhat64



Joined: 14 Mar 2008
Posts: 246
Location: USA

PostPosted: Wed Mar 04, 2009 7:32 am    Post subject: Reply with quote

Paul,
Good idea editting your Page 1 entry. I may get rid of the clutter and edit my original post to point down to yours.

NOTE: Your Syntax leads one to believe you can enter "15%" as a threshold value. Because the percent sign is a special character you MUST enter it as "15%%" or ".15". Otherwise the "%" is ignored and the value is interpreted as a physical threshold of 15GB.

Of course your comment about the awkward If,Elseif,Else iterations got me to thinking. I've also been reading up on VBScript/WMI optimization techniques and I've got a concept that takes on both of those issues (I'd love for this to be a bit faster, wouldn't you?).

I'm still noodling on it, and I'll only post new code if it's a more elegant solution with a performance benefit. I'm slammed for the rest of this week, but stay tuned for future developments.
Back to top
View user's profile Send private message Send e-mail
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Thu Mar 05, 2009 3:16 am    Post subject: Reply with quote

greyhat64 wrote:
NOTE: Your Syntax leads one to believe you can enter "15%" as a threshold value.
This is the case with my test set up - maybe it is because I run a Shell Script??
Parameters: MHDC08 C:2 20%
Response: C:3.1(31%), D:71%(4.28), E:98%(57.49)

cheers, Paul
Back to top
View user's profile Send private message
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Sun Mar 08, 2009 4:15 am    Post subject: Reply with quote

A little bit of code tidying in the parse function to remove superfluous "if ParseOK" tests.

cheers, Paul
Back to top
View user's profile Send private message
greyhat64



Joined: 14 Mar 2008
Posts: 246
Location: USA

PostPosted: Mon Mar 09, 2009 3:39 pm    Post subject: Reply with quote

Good updates!
I'm still noodling out the whole %% thing. Maybe you're right about the shell script thing - I haven't had a chance to look too hard at it.

In truth, I'm having too much fun doing a total rewrite.
I've already found an interesting way to parse arguments that I should've thought of long before, not to mention a more streamlined method of error handling. Now I'm working on a different way to approach the five different 'Test types' which should reduce lines of code significantly, assuming I haven't forgotten something stupid, like error handling.
I'm almost ready, and I plan to 'flood test' my new script against 30-40 machines. I'll only post it (in place of the original code posted) if it performs better in a heavy load situation. We'll see. Right now I'm just having fun learning some new coding methods!

Stay tuned.
Back to top
View user's profile Send private message Send e-mail
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Tue Apr 07, 2009 8:22 am    Post subject: Reply with quote

Now that we have a string of disk tests running I have found a problem if the machine being queried has a wayward WMI service. In our example the WMI service seems to be hung and has to be killed before it can be re-started. This causes the disk test to hang and we now have dozens of "cscript" process running on the HostMonitor server - most of these processes are days old. The workaround for the tme being is to kill these "hung" processes, but doing this manually is painful. I have thrown together a bit of code to kill these processes, that can be run from a test in HM.

cheers, Paul

  1. Select the "Shell Script" test method.
  2. Click on the "Script Manager" icon.
  3. Select "New" for a new script.
  4. Name the script appropriately.
  5. Check the "Start Command" is set to: cmd /c cscript /B /E:VBScript %Script% %Params%
  6. Copy and paste the VBS file into the script box.
  7. Add a hint, e.g. No parameters required, change code to set values.


Code:
Option Explicit
Dim objWMIService, objProcess, colProcess
Dim strComputer, strProcessKill, strHoursToLive, strProcess
Dim strMonth, strDay, strHour, strMinute, strSecond, strDate, strResult
Dim dDate
'## User deined values ##
strComputer = "."
strProcess = "cscript.exe"
strHoursToLive = 8
'## End user defined values ##

const statusAlive     ="scriptRes:Host is alive:"
const statusDead    ="scriptRes:No answer:"
const statusUnknown   ="scriptRes:Unknown:"
const statusUnknownHost ="scriptRes:Unknown host:"
const statusOk      ="scriptRes:Ok:"
const statusBad     ="scriptRes:Bad:"
const statusBadContents ="scriptRes:Bad contents:"

dDate = Now()

' Set date format to allow comparison
If Len(Month(dDate)) = 1 Then
   strMonth = 0 & Month(dDate)
Else
   strMonth = Month(dDate)
End If
If Len(Day(dDate)) = 1 Then
   strDay = 0 & Day(dDate)
Else
   strDay = Day(dDate)
End If
If Len(Hour(dDate)) = 1 Then
   strHour = 0 & Hour(dDate)
Else
   strHour = Hour(dDate)
End If
If Len(Minute(dDate)) = 1 Then
   strMinute = 0 & Minute(dDate)
Else
   strMinute = Minute(dDate)
End If
If Len(Second(dDate)) = 1 Then
   strSecond = 0 & Second(dDate)
Else
   strSecond = Second(dDate)
End If
strDate = Year(dDate) & strMonth & strDay & strHour & strMinute & strSecond

Set objWMIService = GetObject("winmgmts:" _
& "{impersonationLevel=impersonate}!\\" _
& strComputer & "\root\cimv2")

Set colProcess = objWMIService.ExecQuery ("Select * from Win32_Process Where Caption='" & strProcess & "'")
For Each objProcess in colProcess
   if ConvertDateToNumber(strDate) - ConvertDateToNumber(Left(objProcess.CreationDate,14)) _
   > strHoursToLive/24 Then 'Process has been running for over X hours
      objProcess.Terminate()
      strResult = strResult & "Terminated process " & objProcess.Name & ", PID," & objProcess.Handle & " "
   End If
Next
if strResult = "" then strResult = "No processes to kill"
strResult = statusOk & strResult
wscript.stdout.write strResult

' ########## End of script ###########

Function ConvertDateToNumber(strConvert)
' Converts a string in the form yyyymmddhhmmss to yymmdd.part_of_day_as_decimal
' Dates can now be added or subtracted
   Dim sHr, SMin, sSec
   
   sSec = Mid(strConvert, 13, 2)
   sMin = Mid(strConvert, 11, 2)
   sHr = Mid(strConvert, 9, 2)
   if sSec > 0 then sSec = sSec / 86400
   if sMin > 0 then sMin = sMin / 3600
   if sHr > 0 then sHr = sHr / 24

   ConvertDateToNumber = Left(strConvert, 8) + sHr + sMin + sSec
End Function 'ConvertDateToNumber

[Edit 23/7/2009] Error in Month checking. "str = Month(dDate)" should be "strMonth = Month(dDate)".
Not a problem until October.


Last edited by Paul_NHS on Mon Aug 03, 2009 7:51 am; edited 2 times in total
Back to top
View user's profile Send private message
greyhat64



Joined: 14 Mar 2008
Posts: 246
Location: USA

PostPosted: Tue Apr 07, 2009 2:00 pm    Post subject: Reply with quote

Paul,
A nice solution to a sticky problem!
Two thoughts:
1. Maybe a master/dependant relationship should be made between this script and any WMI dependant shell script.
2. Or you could make this a called function of any WMI script you develop.

On a bigger scale, shouldn't AHM either:
XXXXX(a) Timeout: Maybe the Shell Script method should have a default timeout which flag 'Unknown' any unresponsive script.
XXXXX(b) AutoCorrect: Identify when a system has become unresponsive to WMI (or any other) 'shell' request and launch a specified corrective action.
XXXXXAlex, maybe you would like to chime in here

Any clue as to why WMI is hanging in the first place? And don't tell me it's because it's a M$ API - we all know that
Back to top
View user's profile Send private message Send e-mail
Paul_NHS



Joined: 25 Feb 2009
Posts: 59

PostPosted: Wed Apr 08, 2009 1:49 am    Post subject: Reply with quote

Joel,

1. Yes, but it was a quick and dirty....
2. If the WMI script stalls I don't think it will self clean, so it may need to be left stand alone.

(a) We do get "Unknown" in response to the timeout, which I have modified to turn into a warning status.
('%SuggestedStatus%'=='Unknown') and ((%SuggestedRecurrences%==4) or ('%Status%'=='Warning'))
(b) See my solution to (a)

The problem seems to be a stalled SMS Agent Host service (ccmexec.exe). I have to kill both processes before either will re-start. Should have some time to work on it next week.

cheers, Paul
Back to top
View user's profile Send private message
greyhat64



Joined: 14 Mar 2008
Posts: 246
Location: USA

PostPosted: Wed Apr 08, 2009 11:54 am    Post subject: Reply with quote

2. True, unless you call the 'clean up' function before the WMI call. The problem with this 'solution' is that it's frequently an unneccessary call, just taking up clock cycles. I'm not a fan of unnecessary calls.
(a) So, if an 'Unknown' status is returned, you should be able to create a test using your Shell Script and make it a dependant of all other WMI shell script tests with the "Perform test when Master test has a 'Dead' or 'Unknown' status" criteria.
On second thought, the question then is how to pass the hostname/IP of the unresponsive host to your script. UGH!

As for problems between the SMS Agent Host service and WMI, there are a number of references to problems stemming from M$ updates and certain antivirus products. Good luck!
Back to top
View user's profile Send private message Send e-mail
jivetolkein



Joined: 19 Jul 2007
Posts: 96

PostPosted: Mon Aug 03, 2009 6:57 am    Post subject: Reply with quote

This is some great work here

I never looked any further into this, as HM was originally an interim solution (which has now grown over 700 nodes ^^), each of our sites had agreed SLTs, and the volume level warnings were set here (so I'd no mileage in individual volume thresholds, other than ignoring some), and the output goes into Oracle for logging and reporting purposes, so I didn't think about tying it in with the report generator. I was going to revisit this sometime after the immediate fires had been damped down a bit.

Now I won't have too, cheers!

Had come across the cscript issues too, though not too badly as we tend to use two RMAs per site for Windows servers - the most common issue was servers being decommissioned locally without any warning, so backlog of checks built up each in their own RMA agent server. We had some local checks and kills running on the depots, but I'd never thought of putting it in HM, don't know why!?

Anyway, my main point of posting is - who thinks this should now move to a feature request? Move Paul/greyhats work into an official HM test? I like scripts, but I think this is core requirement for most Windows shops, that perhaps people are missing if they don't follow the forums, and maybe discounting the product without investigating thoroughly enough.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    KS-Soft Forum Index -> Library All times are GMT - 6 Hours
Goto page Previous  1, 2, 3, 4  Next
Page 2 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group

KS-Soft Forum Index