Community discussions

MikroTik App
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

What are your expiriences with "The dude" in produ

Wed Dec 06, 2006 1:50 pm

Hi all,

after I saw the dude several month ago we decided to move the alerting from WhatsUp Gold to this system.

Since three days the system is in live operation and I see many "wrong alerts" for the ping probe. The dude is telling me that the ping probe is down but when you ping the IP the IP is reachable. This happens many times. A parallel ping from the same machine to the same target shows a packet loss of 0% (and no single ping lost).

If I use the ping tool of the dude it will tell me otherwise that the rrt is 1ms. This is impossible cause the target is 3 wireless hops away.

Has someone the same problems?
 
User avatar
antispam
Frequent Visitor
Frequent Visitor
Posts: 62
Joined: Mon Apr 11, 2005 5:57 pm

Wed Dec 06, 2006 4:29 pm

Installing the dude 2.0 made some of our 3Com SuperStack 4400 switch behave very weird. The switches were resetting every x seconds. So a lot of users lost connectivity. :evil: :evil: :evil:
 
User avatar
normis
MikroTik Support
MikroTik Support
Posts: 24605
Joined: Fri May 28, 2004 11:04 am
Location: Riga, Latvia

Wed Dec 06, 2006 4:31 pm

you can configure the ping probe to fail only for example 3 non returned pings. it's highly customizable
 
User avatar
normis
MikroTik Support
MikroTik Support
Posts: 24605
Joined: Fri May 28, 2004 11:04 am
Location: Riga, Latvia

Wed Dec 06, 2006 4:31 pm

Installing the dude 2.0 made some of our 3Com SuperStack 4400 switch behave very weird. The switches were resetting every x seconds. So a lot of users lost connectivity. :evil: :evil: :evil:
tell dude not to poll them. or remove SNMP monitoring for this kind of device.
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Wed Dec 06, 2006 6:21 pm

Normis, I am playing with the dude some month long.
I configured the probe time every 10 second and it should be down after 6 polls. Timeout is set to one second. I think this is not the problem. As I said a ping from same machine runs without a lost one. And the round trip of 1ms is unrealistic too.
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Wed Dec 06, 2006 6:38 pm

And just another nice phenomen:

From one second to another serveral device in different parts of the network went down. If you wait to next poll the stay down. If you ping the devices with ping probe they are reachable. The devices came back when I did a manual reprobe...
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Wed Dec 06, 2006 8:01 pm

Believe it. Device is down but ping from server in running. Time 5 minutes then manual reprobe and device comes up.

Probe interval 30 sec.
Probe Down Count 4 = 2 min

Image[/img]
 
miko5881
just joined
Posts: 7
Joined: Thu Oct 26, 2006 12:09 am

Thu Dec 07, 2006 12:35 am

We just started to use the dude in our District and other then a few minor glitches which are for the most part all corrected it is awsome. Before we used to wait for somthing to go down and for someone to complain then we just opened up a window and did a ping. Now I use the dude even when there is no network issue just for the fact it has all of our devices in it and I made custom tools so I just right click the device and the tool I want to run. We have experianced no stability problems and are SNMP monitoring windows, Linux, alcatel and Sonicwall systems. The only thing that i think I can really ask for more is the ability to script more stuff and the ability to store values in the database and not just avarages.
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Thu Dec 07, 2006 10:45 am

Hmm! In the night I decided to switch dude off...
At the morning I increased the timeout of all probes to 00:00:05 (I think that means 5 sec) and the massive false errors are gone. Very strange! The ping values are all <30 ms.

@MT: Please check the timeout value. Maybe there is an error.

Dude is running on Windows Server 2003
 
User avatar
znet
Member Candidate
Member Candidate
Posts: 134
Joined: Mon Jul 24, 2006 8:07 pm
Location: Houston, Texas

Production Experiences

Thu Dec 07, 2006 1:10 pm

Since you asked...
I have managed to figure out most of everything. Considering the price performance value, it does very well without any documentation. It is not possible to RTFM and learn anything. Nevertheless, by probing around and trying things for weeks, the 2.0 production version is doing quite well for itself. Web server working is a nice touch. SNMP seems to be working.

Minor problems:
1) Sometimes a new map just is forever suspended in the 'Loading' condition. I recommend frequent backups since you will likely soon get to depend on it.
2) No way to do database update forces you to think carefully about major changes. If you go too far and it gets 'stuck', you will lose a lot of work. Only experience can avert disaster there...
3) No Cut and Paste! It is a real pain not to be able to select a device and copy and paste it. This is a especially necessary because of the VERY inconsistent behavior of discovery. Its easier to just manually enter everything, even without cut and paste.
4) Need standard scripting functions like trim() to just use x amount of characters in a string. I cant figure out what those built-in functions are doing with 'if string_size'. Maybe I will figure it out...
5) Need more ariithmetic functions. Can divide two results of an oid, and it returns a decimal number with about 5 places after the decimal. No way I have found to truncate that. However, if you try to divide by a decimal number, forget it, that doesnt fly. You must use algebra to determine how to use integers until the last formula! Thats a good trick.
Concatenate seems to be used in strange ways, but with no doc, its trial and error. Also need a way to format a table from the output of oid_column to create a proper table.
6) There is no 'For/Next' functions. You can determine the array size, but what can you do with that? You cant loop through, increment the array element number until you get to the end. I am sure its coming. Just part of the 'price/performance ratio!'....
7) Need to be able to add (sum) two built-in rate outputs. I need to add the bandwidth of two ports since I have an HSRP connection with two Fast Ethernets. Traffic goes over both, and havent figured out how to add the two rates together to chart the aggregate bandwidth. Now I just have to add them in my head.
8 ) Need formatiing capability in the device labels. It just center justifies everything. Some oid results produce variable output and would look better if they were left justified.
9) Need to be able to have a custom main page for user groups. Web interface should be tailored to the audience. Some users just dont need to see certain things, and it just causes trouble when they start looking around in there.
10) Where are the graphs (charts) in the web version They are in the RouterOS.
11) Need current values indicated in graph outputs to show current values like other typical bandwidth graphing tools.
12) Need to be able to draw vector objects on maps. Would like to draw a box around set of devices. Dont see how that can be done.

Well, enough said. This product does things that no other product can do and I am sure that given the rapid development already proven, I am confident there will be all these things and more, soon enough.

Thanks for the product Mikrotik! :D :D
Znetworks
Houston, Texas
 
uldis
MikroTik Support
MikroTik Support
Posts: 3439
Joined: Mon May 31, 2004 2:55 pm

Thu Dec 07, 2006 1:11 pm

complete2006, maybe you should try to adjust ping probes packet size?
Or at least make sure that you are using the same size everywhere.. We have seen problems with ping sizes 28-29
 
mmtc
just joined
Posts: 8
Joined: Thu Dec 07, 2006 12:38 am

Thu Dec 07, 2006 9:23 pm

miko5881: What kind of custom tools do you have? Would you mind sharing some of those custom tools, please?

:D
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Fri Dec 08, 2006 1:09 pm

Thank you uldis for your response. I changed the ping size from dudes default to 32 bytes at the evening. In the morning there was no SMS and I thought: That did the job. Now I know that the GSM gateway from 2N was crashed.

So I will tell here at the evening if the ping probe size changement will help. Yet I can say that the problem of down device while the device is reachable is still there. (see above)
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Fri Dec 08, 2006 7:26 pm

@Uldis: The ping of size 32 Byte did the job. No mass-alarm-sms till now.
Can you explain why the default values will not work (mosty ROS in the network)?

[/quote]
 
complete2006
Member Candidate
Member Candidate
Topic Author
Posts: 235
Joined: Tue Feb 07, 2006 7:18 pm

Mon Dec 11, 2006 10:28 am

Update:

The 32 Byte did only the half job.

My question: If I put a retry value of 5 in the ping probe will the probe then alway ping 5 times or will it ping upon the first response of the device and stop then?
 
komandor
just joined
Posts: 1
Joined: Thu Jun 15, 2006 2:56 pm

Mon Dec 11, 2006 12:37 pm

I have the same problem with ping from Dude.
In most cases with Zyxel end other shdsl modem.
Some times Dude informs that the device is not accessible
But from the Unix console the device normally answers on icmp.
From unix 0% lost
From Dude 100% lost

Dude runs on Windows XP Proffessional SP2
Can for Deude the server platform is necessary?
 
miko5881
just joined
Posts: 7
Joined: Thu Oct 26, 2006 12:09 am

Mon Dec 11, 2006 6:12 pm

mmtc: The tools I was talking about are things that I can add in the tools list and associate with a device type. For example for all of our windows computers I use remote desktop so I added a tool that I can run the remote desktop client and connect to each one just by right clicking. Our novell servers have between 2-3 web serves running on them so I add tools for each one. I added ssh so I can just ssh to any monitored linux system. I use putty which is a small and free download (http://www.chiark.greenend.org.uk/~sgtatham/putty/). I have ldap tools that I run on some of my systems. Also I was playing around with the idea of a shutdown/reboot/WOL tool. It will basically be a script that can shutdown or reboot and if I have the mac address power it on. I already do that in some of my scripts but I have not attempted it in the dude yet. I think what could make it even better is if there was a way to code GUI stuff right into the dude so for my shutdown/reboot script I can have a GUI come up built into the dude and have code behind it that will connect and run commands on the server. But for now being able to use 3rd party programs like that makes it much nicer then having to remember all the programs and then having to type in the host names or ips of all the system I want to connect to. Sorry if that response was long but let me know if you want more info.
 
mueller
just joined
Posts: 17
Joined: Mon Nov 20, 2006 7:01 pm

Mon Dec 18, 2006 1:21 pm

I have been using the dude for about 4 month and have to say it is one of the best tools on the net. I have a ton of hp openview NNM experience on large networks and have also used many other management tools.
Compared to the dude are are some major advantages to the high priced solutions but they are high priced and often lack in custamizabilty.
I recommend the dude to everyone I work with.

I have rolled it out on several networks including 1 1000+ switch networks and so far have had only minor problems.
 
User avatar
jphilbert
just joined
Posts: 13
Joined: Fri Sep 22, 2006 5:22 am

Dude Experiance

Wed Dec 20, 2006 3:58 pm

I have been using dude for over several months. for my IT education department its and invaluable tool.

Only thing I like to see is, better reporting and more detail graphs and some kind of SQL support.
 
sten
Forum Veteran
Forum Veteran
Posts: 920
Joined: Tue Jun 01, 2004 12:10 pm

Wed Dec 20, 2006 6:27 pm

The Dude is a pretty cool product.

Last time i tried Dude it was quite network resource hungry.
The average packets pr second the application prefers to send is somewhat high. For statistics to be correct and the projected response times to be correct, one must schedule the outbound requests (X) in such a way that there is a delay (Y) between each transmission.

And that rate should be as close to:

Y := ( time interval / X )

If it's bursty then the statistics would be wrong and the application could be seriously detrimental to your network.
Move along. Nothing to see here.

Who is online

Users browsing this forum: No registered users and 15 guests