Jump to content
Welcome to our new Citrix community!
  • 0

Citrix Hypervisor Time offset meaning


Kris Radke1709156948

Question

I am trying to figure out a vda registration issue and have banged my head against most everything, and have comp to an oddity that I am missing some understanding with XenServer basics regarding time.  My customer came to me just after an outage. 

Env overview. 

VCAD is 7.16 with VDA 7.18 currently.

XS 7.6 for my VDA pool 

XS 7.3 for my Server pool

 

I have cleaned up a lot of bad in AD and I think fixed a time sync issues as well. or so I think. 

So the VM in question is showing HostAgent = LOCAL for time.  Windows time across VDA, DDC and DC match. So I should be good. 

 

I look at the host that the VM resides on and run ntpq -p

     remote           refid                          st t when poll reach   delay   offset  jitter
==============================================================================
 "DomainConntroller1"  .LOCL.           1 u   27   64  377    0.436  -8866.5  12.744
 "DomainController3"   "IP of DC1"     2 u    5   64  377    0.614  -8864.6  12.256

 

Host that my DDCs reside

 

     remote           refid                       st t when poll reach   delay      offset  jitter
==============================================================================
 104-168-88-15-h 164.67.62.194    2 u  76d 1024    0   58.526   -0.168   0.000
*lga1.m-d.net    204.9.54.119     2 u  984 1024  377   33.443   -1.376   1.087
 

AND

 

    remote           refid                         st t when poll reach   delay     offset  jitter
==============================================================================
 "DomainConntroller1"  .LOCL.           1 u   25   64  377    0.495  -9222.5   6.914
"DomainController3"   "IP of DC1"     2 u   49   64  377    0.947  -9226.6   5.000

 

And I have another host that is like so

 

   remote           refid                             st t when poll reach   delay   offset           jitter
==============================================================================
 "DomainConntroller1"  .LOCL.           1 u   20   64  377    1.185  22664.3  13.301
 "DomainController3"   "IP of DC1"    2 u   12   64  377    0.923  22658.0   9.704
 

 

 

I also looked at the time offsets of the VMs themselves:

All of them are within a few numbers of each other

name-label ( RW)    : Test020
      platform (MRW): timeoffset: -18007;

 

DDC1 DDC2 and DC1 and DC3 all were -18000 

 

Which I would believe is UTC -5 Eastern timezone. 

 

I do see that one server is using Public NTP, and those addresses are what my PDC "DC1" use as well.

With me being n00bish in the XS world. What am I missing? 

Could it be the gross offset from host to host? I see one is 22664.3 and the other -9222.5?

What does that mean? 

Is that the cause for my VM to register almost 11.5 hrs later? 

 

Link to comment

10 answers to this question

Recommended Posts

  • 0

The offsets are huge in the cases you mention.  Are they all connected to the same NTP server(s) and are their time zones all the same?  If offset by so much, you may need to stop the NTP daemon, manually reset the time pn the server, and re-enable NTP The VMs wil by default get their times from the XenServer host. All systems should be using the same NTP source, both for XS and your VDAs.

 

-=Tobias

Link to comment
  • 0
23 hours ago, Tobias Kreidl said:

The offsets are huge in the cases you mention.  Are they all connected to the same NTP server(s) and are their time zones all the same?  If offset by so much, you may need to stop the NTP daemon, manually reset the time pn the server, and re-enable NTP The VMs wil by default get their times from the XenServer host. All systems should be using the same NTP source, both for XS and your VDAs.

 

-=Tobias

 

Thank you -- In digging in most of the servers are not syncing to any of the NTP servers listed. And since those are VMs inside one of their pools. I am changing the same sources for all host. Matching what is in AD under group policy.  

I have some of the hosts are set to with the HWCLOCK_SYNC=YES and others are not. I have seen in some of the forum post that with the Dell R series hosts there are some issues with time sync. Right now I cannot upgrade past 7.6 due to licensing. Do you have any thoughts on which way to set. I am reading through your "Tips" post right now to better understand the 'bester' way of setting up NTP. It maybe answered in there, I haven't made it all the way through yet. 

 

Again Thank you. 

Link to comment
  • 0

Yes, check out the tips articles and make sure to check some of the comments later on as things keep changing from release to release, in particular whether to make use of the HW clock or not. We use internal NTP servers, which in turn sync to external sources, the advantage being outside connectivity can fail and you still have access to a common NTP source locally. Good luck!

 

-=Tobias

Link to comment
  • 0
1 minute ago, Tobias Kreidl said:

Yes, check out the tips articles and make sure to check some of the comments later on as things keep changing from release to release, in particular whether to make use of the HW clock or not. We use internal NTP servers, which in turn sync to external sources, the advantage being outside connectivity can fail and you still have access to a common NTP source locally. Good luck!

 

-=Tobias

Gotcha, That is why I had them set to two DCs first however they are failing to sync for some reason.

Since I am not as experienced with XS in general. I plan on adding external sources and keeping the two internals having to 4-6 entries and make sure there is an authoritative source that the hosts are syncing to. Then try to figure out why they are not using the internal DCs to sync. In their lab pool.

It does look like the host that is referencing the two external is does not have a gross offset. I will start with those.  

Link to comment
  • 0
On 2/10/2020 at 11:17 AM, Tobias Kreidl said:

Sounds encouraging ...keep us posted!

So still dealing with some sort of issue from within AD for NTP.  I have not found anything. But I cannot get my PDC role DC to sync externally out from the hosts. Keep pointing to Local CMOS. The customer mentioned that he read somewhere that there was some issue with a VM not being able to from XS hosts. I did check the AGENT HOST registry keys. both locations are blank. I was thinking as long as those key values were not  "UTC" it would not use the hardware clock. 

 

Link to comment
  • 0
46 minutes ago, Kris Radke1709156948 said:

So still dealing with some sort of issue from within AD for NTP.  I have not found anything. But I cannot get my PDC role DC to sync externally out from the hosts. Keep pointing to Local CMOS. The customer mentioned that he read somewhere that there was some issue with a VM not being able to from XS hosts. I did check the AGENT HOST registry keys. both locations are blank. I was thinking as long as those key values were not  "UTC" it would not use the hardware clock. 

 

Correction - It is now corrected and I retract my previous question. --- I am waiting now for all to propagate out. Thank you!

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...