RTC problems on Dell Poweredge Servers

[CIDM]

Hi All !

This one is a bit technical and i hope we come to a solution.

I finally managed to install a VMWare server on Debien Etch kernel 2.6.18-3-686. Beside all the problems noted and fixed in another post (http://www.howtoforge.com/forums/showthread.php?t=9429&highlight=dell) I still have an issue related to RTC.

This is what should happen on a normal system :

[root@testsystem ~] hwclock
Wed 03 Jan 2007 06:07:26 PM EST -0.758061 seconds

[root@testsytem ~] cat /proc/interrupts
CPU0
(snip)
8: 8 IO-APIC-edge rtc
(snip)


The number 8 would increase beside IO-APIC-edge rtc showing that hwclock receives correct interrupts form /dev/rtc...

That's not the case on the Dell PowerEdge 860.

[root@pe860 ~] hwclock
select() to /dev/rtc to wait for clock tick timed out

[root@pe860 ~] cat /proc/interrupts
CPU0 CPU1
(snip)
8: 0 0 IO-APIC-edge rtc
(snip)


This issue is corrected on Dell servers by adding --directisa option to a hwclock command or in /etc/init.d/hwclock.sh for booting slowdowns and shutdowns... But the issue is affecting directly VMware server.

This message is logged in syslog


Jan 3 16:08:41 pe860 kernel: rtc: lost some interrupts at 2048Hz.
Jan 3 16:09:12 pe860 last message repeated 1512 times
Jan 3 16:10:13 pe860 last message repeated 3051 times


The final issue is that the time on the virtual machine is sooooo slow. 1 sec = 5 sec realtime. This is I belive because /dev/rtc is broken and must be fixed in the kernel (or rtc module)

At this point, any ideas are welcome.

Thanks !

Francis

[falko]

Are you talking about the host or the guest system?

[CIDM]

Hi Falko,

I'm talking about the host system. It generates errors in syslog because a guest system needs the host's RTC to overcome the 1000 hz or so clock tick limit.

I've tried to modprobe -r rtc , modprobe rtc-dev...

hwclock works fine, but there is no /dev/rtc anymore in /proc/interrupts... this way, vmware is not able to cope with the rtc

Beside that, haven't found anything yet !


Francis

[falko]

And this happens only on a Dell server, not on other servers?

[CIDM]

Hi Falko,

Thanks for your time in this...

I did my homework : It tried the same command (hwclock) on at least 8 of my Debian (sarge) servers. On these I have 2 Poweredge 430sc and 1 Poweredge 860. All of them show the behavior, no rtc interrupts possible. Even though I saw Dell was ok with Debian, having only problems with network and sas drivers. (wich are ok in etch).

On the rest of the servers, they are all dffirents machine (an old IBM, some clones, on AMD + MSI mobo) they all have correct results regarding RTC.

I'll post anything I find...


Francis

[falko]

Maybe these links help: http://wiki.osuosl.org/display/LNX/Debian+on+Dell+Servers
http://linux.dell.com/distributions.shtml

[CIDM]

Hi Falko !

Finally the results of my adventures in RTC :) The debian based setup I use to have (debian sarge 3.1 and debian etch) were all wrong when rtc was needed. I finally moved the whole setup to Ubuntu (your perfect setup). It all went perfectly well on my Dell PE 860...

I traced back the bug I had to the hpet timer function that were not implemented in debian sarge and etch. Having to recompile my kernel leaded to /dev/sda1 not being detect at boot (IDE - SATA order, other problems...argg) So I finally dropped Debian for Ubuntu LTS (not so different since Ubuntu is based on debian)...

I liked Ubuntu very much, and it solved all my worries...

My vmware (Ubuntu) server is now completed. I have Windows 2003 Web ed running + Ubuntu - ISP config running on the other vm. Everything looks cool !

Thank you so much for all your help and wonderful control panel.


Francis

[Toontje]

Which Etch-ISO did you use?

I have an AMD64 ISO (I have an AMD 3800+ on an ASUS A8V-VM SE MB), but I have problems installing the linux-headers.

Edit:
I should read better! You moved to Ubuntu....... Sorry:rolleyes:

The problem is that 6.06 LTS doesn't support my Southbridge (VIA VT8237A), so I will try to install 6.10

[CIDM]

Right now my server is working A1. No glitches. I now provide my own DNS on both ISPconfig servers. No bugs at all !

Good Luck with 6.10. Should work like a charm.

[Toontje]

Also installed 6.10 and no problemo at all!

VMware server is running smooth and fine

[edge]

CIDM,

I've got 2 Dell PowerEdge SC1425 with Debian Sarge on it.

Not sure if this is okay, or not, but your commands are showing me this:

host:/home/edge# hwclock
Sun 25 Feb 2007 04:07:10 PM CET -0.274992 seconds



host:/home/edge# cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 148 0 0 2203400325 IO-APIC-edge timer
9: 0 0 0 0 IO-APIC-level acpi
14: 0 0 1 12 IO-APIC-edge ide0
169: 0 0 0 0 IO-APIC-level uhci_hcd
177: 0 0 0 0 IO-APIC-level uhci_hcd
185: 0 0 0 1 IO-APIC-level ehci_hcd
201: 0 0 0 1504553283 IO-APIC-level eth0
209: 0 0 0 178945454 IO-APIC-level 3ware Storage Controller
217: 0 0 0 3258702 IO-APIC-level eth1
NMI: 0 0 0 0
LOC: 2203548572 2203548571 2203548570 2203548569
ERR: 0
MIS: 0
host:/home/edge#