Ise sends replication alarms but all nodes seem to be synced

[marco.merlo]

Hi,

sometimes in our 2.3 ISE Deployment replication errors show up in main dashboard 

Here is the last one repeated for each not primary node 

"Error while synchronizing EDF objects. Please re-sync the node Please re-sync the node".

 Anyway in System->Deployment all nodes are marked as connected with  "Sync Status:
0 messages to be synced."

Should the node be re-synced or not?

As far as I remember re.sync operation make the secondary node to restart and I don't want to reload a node just for e cosmetic issue.

Regards

Marco

[Arne Bier · Accepted solution]

I remember having this issue too and I have never seen it again in ISE 2.4 - I suspect it was a bug of sorts.  We never manually sync'd our nodes.  I may have raised a TAC case on this and they said it was "cosmetic".  Rather err on the side of caution and open a TAC case.  It's not an error that any customer should have to live with (or see, in the case of a "cosmetic" issues). 

BTW: 2.3 was a terrible release.  i'd make plans to move to 2.4 :)

[marco.merlo]

Hi,

thank you for the reply. Unfortunately at the time we deployed ISE 2.3 was the only version available with the all the features we need. Actually it has lots of annoying (often "cosmetic") bugs on reporting side.

Hoping that 2.4 won't be worse...

Regards

MM  

[jm.virtual01]

I am facing the same issue with ISE 2.6

Just up-graded the ISE from 2.2 patch 14 to 2.6 patch 3 but getting this alarm. All nodes looks good and in sync

Is there any how i can confirm that this is the cosmetic alarm only no any issue?

In the replication log i am not seeing any replication stopped alarm from ISE, is there any suggestion to get rid out of this situation?

[marco.merlo]

Hi,

on our deployment we had  faced this issue just when changing alarm settings and in order to get rid of the error messages we had to reload just MNT nodes. In the meantime we upgraded to 2.4 but to be honest I did not try to modify alarm settings since then, so I can't tell you if this "cosmetic bug" has been fixed, even if from what you are telling it seems still  present in 2.6 . Maybe you can try and restart just MNT nodes  .

Regards

M

[Jason Kunst]

---

@marco.merlo wrote:

Hi,

on our deployment we had  faced this issue just when changing alarm settings and in order to get rid of the error messages we had to reload just MNT nodes. In the meantime we upgraded to 2.4 but to be honest I did not try to modify alarm settings since then, so I can't tell you if this "cosmetic bug" has been fixed, even if from what you are telling it seems still  present in 2.6 . Maybe you can try and restart just MNT nodes  .

Regards

M

---

Only the TAC will be able to dig deeper with you, please open a case if you have a concern

[marco.merlo]

At the end I figured out the which configuration Items were not replicated.

They were changes on alarm settings (System-> Settings->Alarm Settings -> Alarm Configuration) and actually in order to see them work I had to re-sync at least PMN node

M

[jm.virtual01]

Is there any guide line to deal with the replication issue on ISE 2.6?

I am getting this alarm but in the deployment it looks like it is false alarm but wanted to confirm it, is there any way to identify this alarm status?

Here is the error message that i got for the replication error,

Replication Failed : Server=madlab03isemnt1; Message=Error synchronizing object: Resource[NAC Group:NAC:CoAType:ProfilerCoAType]; Operation: Update

---------------------------------

Replication Failed : Server=madlab03isepsn2; Message= Error while synchronizing EDF objects. Please re-sync the node

----------------------------------

These two type of error messages i can see inn the alarm but in actual deployment all looks good?

Does anyone has any suggestion on this?

[Jaws]

This type of error:

Replication Failed : Server=madlab03isepsn2; Message= Error while synchronizing EDF objects. Please re-sync the node

was what we saw after we had applied the patch manually to the deployment.

Forcing a manual synchronisation of each node  from the deployment screen- to transfer the whole database - resolves the issue.

[jm.virtual01]

yes that will resolve the issue and we did it but still we are getting the same error continuously. From the GUI and from the logs we can say that all the nodes are in proper sync but still we are getting that error.

I am trying to find a root cause for this issue.

Is there any idea for this issue?

[Minnesotakid]

Adding to the thread - I am still getting this supposed cosmetic error in ISE 2.7 patch 4.

[t3rebello]

Adding to the thread - I am still getting this on ISE 3.2.0.542 patch 2,3

[cappelenx]

+1

[Chris_Schubert]

Seeing this error on ISE 3.3 patch 2.  It's the first weekend with four servers running that version.   The entire set does reboot over the weekend, so it and some other errors may be related to that, but the same servers never reported some of the errors I am seeing this weekend before the upgrade.  Trying to make sense of what the new normal is on this version.

[MaErre21325]

I have the same issue on verson 3.1 patch 8.

In the logs i found these errors:

2024-02-24 17:37:20,148 INFO   [admin-http-pool15][] com.cisco.epm.jms.AQMessgeHandler -::::- Publishing message for event [TxnCommit / commit] and message class[com.cisco.profiler.im.edf.EDF2EndPoint]

2024-02-24 17:37:22,555 ERROR  [admin-http-pool2][] cpm.edf2.remote.impl.RemoteServlet -::::- Host Name is null. Skipping Remoting Count

They might be investigated by the TAC?

We're receiving a lot of alarms althought the nodes are correctly synched.