Slow booting with CentOS 5.1/x86-64. Delay between initrd info and kernel alive messages

[yusufg]

Hi, I just installed CentOS 5.1/x86-64 with the Gnome desktop on my Fusion 1.1.1 on MacOSX 10.5.2

I applied all updates and also installed vmware tools.

Yet I observe that booting is slow and there is a 40-50 second delay between the initrd information

printed out on screen and the next message which says 'kernel alive'

Does anybody have suggestions on fixing this ?

Regards, Yusuf

[dariush]

Additional Information...

We are seeing the same problem on Fusion 1.1.1 on Mac OS 10.5.2 and Workstation 6.0.3 build 80004 on WinXP Pro SP2. In both cases underlying hardware is Core2Duo

Guest OS CentOS 5.1 with 64bit 2.6.18-53.el5 kernel.

CPU/NetworkIO/DiskIO/Memory usage are all negligible, the kernel just appears to pause for around 55 seconds.

Given the early stage in the boot process nothing else really matters (ie it's all in the kernel)

Any hints on places to look for answers would be much appreciated

Dariush

[admin]

This sounds like a known issue where certain 64-bit 2.6.18 and 2.6.22 kernels do some strange things while decompressing the kernel. I'm not sure if there's a suggested solution aside from a guest kernel update.

http://groups.google.com/group/fa.linux.kernel/browse_thread/thread/51b8defe5823535d/2cc96906679cbeb...

Edit: It's fixed internally. As you probably know by now, though, I can't say anything about timelines.

[wolfwolf]

Hi,

I hit the same issue with VMware Fusion 1.1.2 running on a MacBook Pro C2D 2.4Ghz with Mac OS X 10.5.2 and a VM running CentOS 5.1 x86_64 with the latest kernel update.

Any ETA for the fix?

Thanks.

[fm_cpaul]

I still get this delay at boot time. Latest Fusion (1.1.3) and Mac OS X 10.5.4. Latest CentOS kernel (2.6.18-92.1.6.el5). Any chance of an inclusion of the "internal" fix in the public VMWare Fusion 1.1 branch?

[admin]

Comments in the bug suggest it would be nontrivial to get working on the 1.1.x branch (which implies that 2.0 shouldn't have this problem, but I haven't tested). Also, that 2.6.23.8-34.fc7 and later (not sure what CentOS that corresponds to) shouldn't have this problem (we submitted a patch).