cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
4847
Views
9
Helpful
32
Replies

How to fix Fault Tolerance on ACE 20?

netternewbie
Level 1
Level 1

Hi,

 

I have two ACE 20's in two 6500's. I have an Admin context with a number of contexts. Unfortunately one context has the backup as FSM_FT_STATE_STANDBY_COLD not Standby_Hot.

 

I have checked its configuration and noticed a number of lines missing from the correct active context. I know now I have to force the config across by taking the ft-group out of service.

 

My question is do I do this on the working ACE or do it on both ACE's admin groups? 

 

I think this is what I do.

config ft group X

no inservice

Give it a few mins to sync across to broken ACE and then put it back inservice?

 

Or do I need to do this on the broken ACE? I don't want to lose the good config and the broken config to copy across.

32 Replies 32

Hmm..strange. So you don't see that configuration in standby after replication happens? This shouldn't be the case. What ever configuration you do should replicated to standby automatically. Standby cold in fact means that device has same exact configuration but may have different sw version, scripts, certs or license. Can you do no ft auto-sync running-config and ft auto-sync running-config and see if that replicates configuration? Can you also send me show ft group detail from Admin context. Also, try and get the script matched first and try again.

Regards,

Kanwal

 

Yes very strange, it wiped the config in active and standby ace. It was just a sticky serverfarm. Fault tolerance works but the very minute I put the sticky serverfarm on active config the standby goes to cold and it doesn't have config.

If I break ft and force syn it symcs but delets config. Very strange. I will try and get script matched first. How do I copy it off disk0 to tftp server?

 

Then do I go on standby and delete what is there and copy in script from tftp server?

Hi,

Is that serverfarm associated with that scripted probe?  if yes, try removing just the probe and see if it works fine? Yes you can copy the script to ftp server, and then put in standby and then delete the one which you don't need to standby and try again.

 

Kanwal

Hi Kanwalsi, 

Thats the funny thing the scripted probe isn't associated with the serverfarm where the config gets deleted. This is relatively new  config section.

I am not sure if it is the scripted probe at all since it seems to be working fine now. Its just when I add in this stickyserverfarm it seems to break. I will add in a test probe tomorrow and just see does that sync across.

When the standby is active cold it doesn't have the new stickyfarm config. I was wondering could I manly put it in at this stage on the standby config?

Hi Netter,

This is strange because if it is removing serverfarm configuration, it must be removing policy etc associated with it. You cannot simply delete a serverfarm unless it is NOT associated with any policy. If it is standalone then that's another thing. Can you do "show ft config-errors" and see what error do you get while that serverfarm is in the config? Can you also get me show ft history cfg_cntlr output?

Regards,

Kanwal

Hi Kanwalsi,

 

Firstly thanks for all your help on this. When I first noticed standby cold I checked both config and noticed the below context was in active config but not standby config:

sticky ip-netmask 255.255.255.255 address source STICKY-SSL-****-FARM
timeout 720
timeout activeconns
replicate sticky
serverfarm ****-FARM

 

policy-map type loadbalance first-match ****-HTTPS-POLICY
class SSLCLASS
sticky-serverfarm STICKY-SSL-****-FARM
ssl-proxy client SSL_CLIENT
class class-default
sticky-serverfarm STICKY-SSL-****-FARM
ssl-proxy client SSL_CLIENT

 

When I removed this I could get them to sync. The very minute I add the sticky farm it goes standby cold.  

 sh ft config-error 
No bulk config apply errors

 

I have uploaded  show ft history cfg_cntlr output   in attached file

Hi Kanwalsi,

Sorry discovered this from standby config:

 sh ft config-error 
Wed Jun  4 11:40:39 UTC 2014
 
`sticky ip-netmask 255.255.255.255 address source STICKY-SSL-****-FARM`
Error: sticky resource not available
--
 
`sticky-serverfarm STICKY-SSL-****-FARM`
Error: Sticky group does not exist!
--
 
`sticky-serverfarm STICKY-SSL-FILR-FARM`
Error: Sticky group does not exist!
 
Error(s) while applying config.
 

 

Hi Netter,

If you haven't allocated any resource for sticky, please do so and try again.

Regards,

Kanwal

Thanks Kanwalsi,

I have this on admin context on active ace:

resource-class *****
  limit-resource all minimum 0.00 maximum unlimited
  limit-resource sticky minimum 10.00 maximum equal-to-min

Can I add it to admin context on standby ace without any interference to services in that context or other contexts?

Hi Netter,

Yes it should be fine. The allocation of resources can be done dynamically and its effect will take place once resources are freed. But there would be no impact.

Regards,

Kanwal

Thanks,

Just did that but still in cold state after trying to force a sync. Will try again.

Hi Kanwalsi,

 

I managed to get  it in hotstandy state by removing sticky farm config from active one and then sinking across.

But the very minute I add in sticky farm it goes to cold.

Any ideas?

Hi Kanwalsi,

Just another update I took out the sticky farm and go hot standby working. I just added a test probe and it sync'd across fine. Just seems to have issues syncing across the sticky serverfarm although the resource error doesn't appear on standby one now.

Hi Kanwalsi,

Here is a copy of the sh ft history cfg_cntlr on the admin context of active ace and standby ace after I add the sticky serverfarm config.

Active ACE:

78:1958 => Jun 04 14:30:47: cfgcntlr_handle_mts_msg:4221 Received
startup cfg changed for context 0
79:1959 => Jun 04 14:32:36: cfgcntlr_handle_mts_msg:4143 Received
STOP_ALL_SYNC notification for FT Group 7
 End of Debug History Log

Standby ACE
62:1192 => Jun 04 14:30:24: cfgcntlr_handle_mts_msg:4221 Received
startup cfg changed for context 0
62:1193 => Jun 04 14:32:36: cfgcntlr_handle_mts_msg:4195 Received
STANDBY_COLD notification for FT Group 7
 End of Debug History Log

 

Hi Netter,

If the resource error is gone, do you get any other error if you do show ft config-error? So as soon as you apply sticky serverfarm, you faced the issue. Please do one thing if okay- Try and remove the sticky serverfarm, sticky group and do the sync. After that put in  the configuration again and see if that makes a difference.

Regards,

Kanwal

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: