Forum Discussion

dragonflymr's avatar
dragonflymr
Icon for Cirrostratus rankCirrostratus
Apr 08, 2015

VE and "Clock advanced by..." in log

Hi,

I have four VEs running on ESX 5.1.0. All VMs are configured with 2 vCPU and 8GB RAM. I have both HA pair running 1.2.0HF7 (LTM, APM, ASM, AVR) and Standalone 1.6.0HF4 (LTM, ASM, AFM, AVR) and other 1.6 Standalone with GTM and AVR only. On all VE except one with GTM I have plenty of messages listed below.

Wed Apr 8 15:02:32 CEST 2015    notice  bigip31 tmm1[15376] 01010029    Clock advanced by 1452 ticks
Wed Apr 8 15:02:32 CEST 2015    notice  bigip31 tmm[15376]  01010029    Clock advanced by 1431 ticks

I was reading SOL and posts but still can't figure out why those messages are present and how to fix VEs to avoid Clock advanced. Those VEs with messages are not at all loaded (this is LAB setup), processor is idling, plenty of RAM free, almost no connections processed.

Somehow (looking at provisioning) it seems to be related to ASM - anyone experienced similar issue?

Piotr

application delivery
cloud
LTM
management
TMOS
vmware
  • amolari's avatar
    amolari
    Icon for Cirrostratus rankCirrostratus
    Apr 08, 2015

    I did, and many times (on Viprion). This is the most absurd SOL F5 ever published (long time ago). After many discussions with different F5 insiders.. the message is related (most of the time) with IO latencies. Therefore one may see it often on a Viprion (non-SSD) with many vCMP guests (IO perf are divided ..)

     

    For example, I've seen this almost all times when importing large CRLs (lot's of IO)

     

    BIGIP/Viprion vCMP capable platforms with HDDs have been relatively quick rolled out with SSD platforms and EOLs issued to the former.

     

    As you run on VE, probably you do not have enough IO (is your VE datastore local?)

     

    Alex