Clock source unstable, performance is abysmal

  • Thread starter Thread starter henrik.beckman
  • Start date Start date
H

henrik.beckman

Guest
[EDIT] corrected verion numbers to the correct ones [/EDIT]

Hi,

We have serious issues after upgrading from 2.4 to 2.5, roughly a week after the upgrade we started getting clock source unstable error messages and performance tanked.
We are running our gateway in vmware 4.0 single CPU and 8GB RAM, the hardware beneath the hypervisor is plenty enough.

We have experimented with various clocksources apart from TSC but to no avail, latest has been the paravirtualizer but we still loose all performance shortly after boot.
Typical boot is normal until proxmox loads, the machine then gets slow and stays that way for atleast a few hours before it usualy gets unresponsive, a reboot in this state takes 45 minutes.

Please advice ASAP, We can provide our customer ID if needed.
For obvious reasons we can communicate through our normal email channel.

Sincerly
/Henrik
 
Last edited by a moderator:
We have serious issues after upgrading from 2.4 to 2.5, roughly a week after the upgrade we started getting clock source unstable error messages and performance tanked.
We are running our gateway in vmware 4.0 single CPU and 8GB RAM, the hardware beneath the hypervisor is plenty enough.

if you upgrade with the servicepack the kernel and vmware tools does not change so the question is what changed as it worked before?
 
Nothing changed since upgrade to crash, increased load only.

We noted that the kernel was 2.6.27-1 and not 2.6.27-2, does that have any significance?

We have now rebuilt a new proxmox from scratch and transfered the config and is now since a few minutes back running on the new machine. Hopefully it will stay stable and snappy, we will then transfer the que from our old host to the new one.

New machine is built from ISO and has 2.6.27-2, we haven installed tools yet.

/Henrik