Learn F5 Technologies, Get Answers & Share Community Solutions Join DevCentral

Filter by:
  • Solution
  • Technology
Answers

F5 LTM 11.5 : SYNC FAILED after applying iapps

After applying iapps we noticed that our system is in sync failed. We tried to resync, offline-online, still in sync failed. Name Sync Status Number of Devices Device Group Type Sync Type DG_ACC_INFRABEL Sync Failed 2 Sync-Failover Auto

Sync Summary Status: Sync Failed Summary: A validation error occurred while syncing to a remote device Details: Sync error on f5mechl2-acc-infrabel.msnet.railb.be: Load failed from f5mechl1-acc-infrabel.msnet.railb.be 01070710:3: Database error (13), Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127. Recommended action: Review the error message and determine corrective action on the device

How can we resync?

3
Rate this Question
Comments on this Question
Comment made 26-Aug-2014 by Noble 6
BIG-IP 11.5.0 Build 4.10.245 Engineering Hotfix HF4 resolved my issue but I had to call in for this release.
0

Answers to this Question

placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I'm having the same problem with 11.5. I just created a f5.dns iApp and now the 2 devices fail to sync correctly.

Sync error on xxxx: Load failed from xxxxx 01070710:3: Database error (13), Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127.

I have tried removing the iApp but the devices will still not sync. I am going to open a support ticket.

2
Comments on this Answer
Comment made 18-Mar-2014 by clesan201305 1
Any answer on the support ticket? Receiving the same issue on 11.5 after creating (and since removing) an iapp. I resolved the issue by changing sync leader and cluster is getting synced, but I receive this error everytime I make a change on one of the nodes, so can only use 1 node to do changes right now.
0
Comment made 18-Mar-2014 by Gordon Johnston 98
Yes, it's a known bug with ID 441512. The way I was advised to temporarily resolve it was to do "touch /service/mcpd/forceload" on the unit not accepting the sync, then rebooting it. Then doing an 'overwrite' sync from the first unit. It's happened a few times since. It seems pushing changes that consist of changes in multiple partitions seems to sometimes trigger it, but I'm far from sure on that.
0
Comment made 13-Apr-2014 by daemien 85
has a hotfix been released yet?
0
Comment made 14-Apr-2014 by Gordon Johnston 98
Not that I'm aware of, I hope it comes soon, I must have hit it about 10 times by now. Hugely frustrating.
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

HF5 is out. It appears to fix this issue:

"474166-2 An error of the form "Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13" will no longer appear."

1
Comments on this Answer
Comment made 02-Oct-2014 by TJ Vreugdenhil 492
Thanks for the update Jie
0
Comment made 02-Oct-2014 by Jie 2658
There is another similar one fixed: "480248-1 Resolved DB 13 error while uploading the UCS."
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I am having this problem as well...

0
Comments on this Answer
Comment made 08-Apr-2014 by shawno 4
I just tore down my device group and rebuilt... fairly simple to do, annoying though.
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

When you run touch /service/mcpd/forceload and reboot second unit that is not getting sync you might lose you last changes that you made on the primary unit

Happened to me for few times. I can't wait for hotfix

0
Comments on this Answer
Comment made 16-Apr-2014 by clesan201305 1
Unfortunately we were informed a hotfix is not coming for this, as F5 has an internally documented valid workaround in the use of set-sync-leader. This does not always resolve it for us. They state it will be fixed in the next major release (12, but could not give an ETA). We have a way to reproduce it in a way that setting config-sync will break the configuration beyond repair of set-sync-leader (link the broken unit to an enterprise manager, create an analytics profile and watch the cluster crumble), and wanted to escalate the case. Unfortunately, F5 requires proof before they will escalate (mainly a qkview after setting sync leader failed), and we cannot continue bringing the customer's cluster down for this, so we have put this on hold for a while. I do not have time to reproduce the issue in a lab environment for at least a few weeks. If someone else does, escalating to get a hotfix would help a lot of people out.
0
Comment made 15-Jun-2014 by daemien 85
is anyone aware if this has been fixed?
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

A temporary workaround was to disable auto sync. Seems to be it solved my issue

0
Comments on this Answer
Comment made 14-Apr-2014 by Gordon Johnston 98
I've had this problem a ton of times now and have never used auto sync. I hope it stays cleared for you.
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I've had this issue since I upgraded to 11.5.1 HF2 och started using iApps. In my case nothing help, it does not matter if I've configured auto sync or manual sync. My own workaround is to restore a .ucs file on the standby node I created after clustering but before creating services. I restore it, restoring does not really work as it should neither so I need to manually reboot it. After that I perform a sync från active (with all the services) to standby (whitout any service). It works any time but it is pretty tedious.

0
Comments on this Answer
Comment made 28-May-2014 by nitass 13347
can you open a support case and ask for engineering hotfix? the bug is tracked as ID441512.
0
Comment made 02-Jul-2014 by Birddog 18
According to release notes, Error 13 is fixed in 11.5.1 HF3
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I'm having the same problem with HF3 installed. :(

0
Comments on this Answer
Comment made 07-Jul-2014 by nitass 13347
so, you had better open a case. :-)
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

I have 11.5.1 with HF3 installed and I have the problem as well. I just opened a case on it.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

About to open a case too, same issue, with HF3.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Lads,

Even though I am not using iApps, the issue has occurred in my Lab as well.

After I created the file forceload in /service/mcpd/ directory, and rebooted the problematic box the issue seems like resolved.

Hotfix is definitely needed though, this is not a simple issue to overlook.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

From release notes of v11.5.1 HF4:

Cumulative fixes from BIG-IP v11.5.1 Hotfix 3 that are included in this release

TMOS Fixes

441512-1

An error of the form "Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13" will no longer appear.

Is this the one? But it has a different error message.

Does this occur on v11.4.1 as well?

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Hi,

I had the same issue.

To fix this I restarted the Device Trust in both units, then I did an upgrade to 11.5.1 HF4 from 11.5.1 (without a Hotfix) and then rebuilt the cluster.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Hi,

also same issue "Database error (13), Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127." on the passive LTM node. Configuration synchronisation was broken.

After the passive device reboot with the mcpd forecereload procedure, the configuration synchronisation between actve and passive was restored but HA was broken (both nodes wanted to be active).

In fact, the port lockdown of my failover interface has switched to "Allow None". Setting the port lockdown back to its previous value has fixed this second issue.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

BIG-IP 11.5.0 Build 4.10.245 Engineering Hotfix HF4 resolved my issue but I had to call in for this release.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Hotfix-BIGIP-11.5.0.4.0.245-HF4 was released. did you mistype that 1 in your build number or is that a patch to HF4 ? I am on HF4 (4.0.245) currently and have not seen the issue recently.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

No. I experienced this issue on 11.5.0 Hotfix 4. Once I was provided the Engineering Hotfix HF4 (4.10.245) I cannot replicate the error.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

There is also a "post" HF4 engineering release for 11.5.1, that I'm still testing.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

We ran into this on 11.5.1 HF4 as well. We need the engineering hotfix soon...

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

Hi everyone, I own active/passive LTM 3900 with 11.5.1 HF4 and I went through the same error ... Except that I never ever used iApp features ... It's not only about Synchronisation, this also happens on the passive device :

[root@eu-dcc-f5i02:Standby:Sync Failed] config # tmsh save sys config
Saving running configuration...
  /config/bigip.conf
  /config/bigip_base.conf
  /config/bigip_user.conf
[root@eu-dcc-f5i02:Standby:Sync Failed] config # tmsh load sys config
Loading system configuration...
  /defaults/asm_base.conf
  /defaults/config_base.conf
  /defaults/low_profile_base.conf
  /defaults/low_security_base.conf
  /defaults/policy_base.conf
  /defaults/wam_base.conf
  /defaults/analytics_base.conf
  /defaults/apm_saml_base.conf
  /defaults/app_template_base.conf
  /defaults/classification_base.conf
  /defaults/daemon.conf
  /defaults/fullarmor_gpo_base.conf
  /defaults/profile_base.conf
  /defaults/sandbox_base.conf
  /defaults/security_base.conf
  /defaults/urldb_base.conf
  /usr/share/monitors/base_monitors.conf
Loading configuration...
  /config/bigip_base.conf
  /config/bigip_user.conf
  /config/bigip.conf
01070710:3: Can't save/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127
Unexpected Error: Loading configuration process failed.

Loading configuration doesn't work at all either.

I opened a support case, I'll let you know what's next

0
Comments on this Answer
Comment made 08-Sep-2014 by Cyril 115
I've been provided an Engineering hotfix that seems to solve the issue ... I guess it will be in HF5 or HF6 edit : build number 11.5.1.4.59.128
0
Comment made 09-Sep-2014 by Jie 2658
What's the build number of this EngHF?
0
Comment made 09-Sep-2014 by Cyril 115
I addded the info in my previous post, sorry
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

FWIW, I saw this issue in 1.5.1, but after applying hotfix 4 I haven't been able to recreate it. I am creating an iApp through a scripted SSH process.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

if it took much time with you for sync, you can go and do it manually, do the same configuration in both nodes then sync.

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

This is ID478690 which was resolved in 11.5.1 HF5 as part of the fix in ID474166.

For more details - see https://support.f5.com/kb/en-us/solutions/public/15000/100/sol15175.html

0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

i'm having this issue with 11.6.0 hf1 - iapps are not used, just started happening out of the blue.

anyone manage to permanently solve this issue?

0
Comments on this Answer
Comment made 21-Nov-2014 by Jie 2658
How did you resolve the issue temporarily? Did you call F5 Support for the confirmation of the issue with a BUG ID? There have been no other hotfixes released for 11.6.0 apart from the hf1 that addresses a single issue of BASH, since 11.6.0 was released on 25 Aug 2014. This might mean, and I can only guess, that 11.6.0 is being abandoned and v11.7.0 is around the corner.
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

i haven't fixed it. i opened a support case, waiting to hear back.

not sure what this has to do with bash, it's a sync issue, this is what i have in the ltm log after attempting the sync:

Nov 21 20:47:40 slot1/xxx notice mcpd[6009]: 01071038:5: Unit key hash from key header: xxx

Nov 21 20:47:40 slot1/xxx notice mcpd[6009]: 01071038:5: Unit key hash computed from read key:xxx

Nov 21 20:47:40 slot1/xxx notice mcpd[6009]: 01071038:5: Unit key read from the hardware.

Nov 21 20:47:40 slot1/xxx notice mcpd[6009]: 01071038:5: Loading keys from the file.

Nov 21 20:47:49 slot1/xxx err mcpd[6009]: 01070710:3: Database error (13), Cannot update_indexes/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127.

Nov 21 20:47:50 slot1/xxx err mcpd[6009]: 01071488:3: Remote transaction for device group /Common/device-group to commit id 20970 6084332014373006884 /Common/xxx 0 failed with error 01070710:3: Database error (13), Cannot update_indexes/checkpoint DB object, class:sflow_http_virtual_data_source status:13 - EdbCfgObj.cpp, line 127..

Nov 21 20:47:50 slot1/xxx err mcpd[6009]: 01071392:3: Background command '/usr/bin/set-rsync-mgmt-fw close' failed. The command exited with status 1.

Nov 21 20:47:54 slot1/xxx notice clusterd[8012]: 013a0006:5: mcpd tells us that config is being saved; incrementing this blade's revision and saving cluster config
0
Comments on this Answer
Comment made 21-Nov-2014 by Jie 2658
You can try the method mentioned earlier in this thread to force mcpd to rebuild its database at reboot, at the standby device, and sync from the active to overwrite what's on the standby. Save your config first as you might lose your last change. as is suggested in earlier posts.
0
placeholder+image
USER ACCEPTED ANSWER & F5 ACCEPTED ANSWER

yeah that's actually what i ended up doing.

# tmsh save sys config
# touch /service/mcpd/forceload
# clsh reboot

after coming back up i forced a sync from the other unit and viola.

clsh is for viprion multi-blade chassis, just regular reboot will work for everything else.

0
Comments on this Answer
Comment made 21-Nov-2014 by Jie 2658
Good to hear this worked for now. Please share the details with us if you do get an engineering hotfix from F5 Support, or even just a BUG ID.
0
Comment made 12-Dec-2014 by Rob 133
Thanks for this solution! I just upgraded to 11.6 HF3 and then installed and removed an external monitor and found I could not sync anymore. This resolved my problem.
0
Comment made 12-Dec-2014 by Nik 270
same error? if the steps i didn't work you can try adding this (run it after tmsh save, before touch) - rm /var/db/mcpdb.*
0