Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..
1419 views

reconciliation.min.ci.rate??

Anyone know what this fuse is for? reconciliation.min.ci.rate 

 


I been seeing this alot in one of our customer environemnt with big enrichment.

 

2013-09-09 19:18:25,706 ERROR - !!!FAILED!!![Operation #1986904278 Name=DataInBulkOperation]  interrupted due to timeout of 4518 sec.
The timeout is calculated according to the minimal data-in rate of 30 CIs/minute.
In order to increase the minimal data-in rate please change the value of "reconciliation.min.ci.rate" in UCMDB settings.
 
When looking for the value of reconciliation.min.ci.rate through JMX console, the value is null.  I couldnt find any documentation or mention of this fuse. 
 
Anyone seeing something similiar? and step to resolve the error? 
 
Thank you
0 Likes
13 Replies
Highlighted
Acclaimed Contributor.. Acclaimed Contributor..
Acclaimed Contributor..

Re: reconciliation.min.ci.rate??

The data-in process in UCMDB is single threaded operation in UCMDB 9.X and 10.X. It means, that when UCMDB is trying to reconcile CI DB is locked. It done for every single CI coming to UCMDB. In case reconciliation operation takes let's say 10 min, any other operations related to create/update on DB level will be blocked. It's not only related to single CIs, but for views, models, and other resources.

The fuse was done to prevent cases where specific CI reconciliation stops all data-in oprations.

IMHO, the way of fixing the problem is to check why reconciliation of specific bulk took so long.

 

I'd like to remind that changing of any fuses without R&D/Support guidelines could cause multiple problems from performance reduction to OutOfMemory situations.

Regards
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.
0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

Looking closer at the reconciliation log I see that when enrichment is run alot of merge operation is also running but no merge actually take place (probably due to bulk fail)

 

One Thing i notice is that when ucmdb is merging the two node it is keeping the old value (old vm name) 

 

Level - 2 - Merging 2 CIs:
Newly Discovery VM - {nt@3cb3e111614174afeea7de50536b35ab}@[(string-bios_uuid:44454C4C-3900-1031-8038-C8C04F504E31), (string-name:dtf04da231cb5c), (string-os_family:windows)(string-serial_number:H918PN1)]

 

Old VM that been disable - {nt@2560846795cf06832b9b87cc2a5b69eb}@[, (string-bios_uuid:44454C4C-3900-1031-8038-C8C04F504E31), (string-name:dixnj-util) (string-os_family:windows),(string-serial_number:H918PN1)]
Merged to:
Kept old VM settign - {nt@2560846795cf06832b9b87cc2a5b69eb}@[string-bios_uuid:44454C4C-3900-1031-8038-C8C04F504E31),(string-name:dixnj-util), (string-serial_number:H918PN1)

 

Since both node are from discovery, I cant set it in reconcialiation priority. How can I set it so that the merge keep the new value of the VM? Instead of the old. 

0 Likes
Highlighted
Acclaimed Contributor.. Acclaimed Contributor..
Acclaimed Contributor..

Re: reconciliation.min.ci.rate??

"Level - 2 - Merging" means that original operation wasn't related to host but to related CI. For instance, discovery set CPU with connected host (string-name:dixnj-util). 

When "HC by *" will send host explicitly reconciliation will do the magic. Please try to do so and let us know.

 

 

Regards
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.
0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

Just to clarify...we should run Host connection by wmi on these windows and it should reconcile and merge the host? 


Can you also clarify what the different level of merge mean Level 1, 2, 3, and 4. 

0 Likes
Highlighted
Acclaimed Contributor.. Acclaimed Contributor..
Acclaimed Contributor..

Re: reconciliation.min.ci.rate??

Level is relationship. For instance you're sending CPU. It should have container to host. So, CPU will be sent with Host.

Merging of CPU will be level 1, when Host will be level 2 etc.

Move reconciliation log to debug and check Level 1 merges.

 

Regards
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.
0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

So I should  run Host connection by wmi on these job and hope that it reconcile? Is that what you mean by "HC by*"

 

Also when I try to manually merge two node, I see alot of merge ooperation level 3 and 4 also happenining with other set of node. Does that mean these merge are for host that are related to a certain CI that are being merge on level 2? (Hope that make sense)

 

 

Does HP have any detail documentation on the reocnciliation process? where it talk about these different phase and level of reconciliation.

 

I read the chapter in the data flow management guide but that is very basic. 

 

Thank you again

0 Likes
Highlighted
Acclaimed Contributor.. Acclaimed Contributor..
Acclaimed Contributor..

Re: reconciliation.min.ci.rate??

I'm attaching Data Flow Management doc from latest version of UCMDB. CHeck the reconciliation chapter it contains more info.

 

We in support are mostly talking problems :).

If you have a problem with reconicliation let us know.

If you would like to get understanding you need to check what was sent and what was recieved.

 

Rerun the discovery, collect the communication log with results. From the log you can see what was sent by the script.

After check reconcilaition.log in debug and you'll see what was on input and what is on output.

 

If after all you think that the problem is in place, please raise support case.

 

 

 

Regards
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.
0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

Running discovery on these node does not seem to get it to reconcile and merge.

 

Discovery would update the new node but would not merge the two. These node attribute are exactly the same except host name and OS version. They share the same IP with different interface, but their interface attribute are exactly the same (name, mac-address, speed, index and alias).  According to reconciliation these should merge but it not. 

 

When I try to manually merge the two, it would merge and keep to old VM attribute and delete the new VM. Is there a way to merge it and force it to use the new VM attribute?

 

Also after the merge, I re-ran discovery and discovery would just recreate the new VM again. 

 

Thank,

0 Likes
Highlighted
Acclaimed Contributor.. Acclaimed Contributor..
Acclaimed Contributor..

Re: reconciliation.min.ci.rate??

I'll need to see reconciliation.log in DEBUG to answer why two instances wasn't merged.

If forum isn't a good place to share such data, please go forward and open support case.

Regards
-Dmitry Gomel, PMP
Click the Like button at the bottom to say 'Thanks'.
0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

The thing is even with DEBUG on, reconciliation.log was not updated after discovery was run. 

0 Likes
Highlighted
Outstanding Contributor.. Outstanding Contributor..
Outstanding Contributor..

Re: reconciliation.min.ci.rate??

So back to my question when reconciliation merge two discvered Node.

 

Wouldnt it make sense to keep the newly discovered attribute? But it seem like ucmdb is deciding to keep all the old value (old host name and operating system)

 

Is there a setting somewhere that specify the reconciliation priority of existing node to be higher than newly discover node? 

Is there a way to set reconciliation priority of new node higher so when it merge it retain the new node atrribute?

 

 

 

0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.