I have VM running under RHEL6, which is running under HyperV. It has a lot of important soft installed, but does not have suffocating workload (~10% CPU is used). It's been really unstable recently unexpectedly showing different I/O errors:
ps aux | grep httpd
-bash: /bin/ps: Input/output error
-bash: /bin/grep: Input/output error
I tried to fix drives by running fsck, but I've got following error:
/dev/mapper/vg_vm-lv_root: recovering journal
fsck.ext4: Bad magic number in super-block while trying to re-opoen /dev/mapper/vg_vm-lv_root
Warning... fsck.ext4 for device /dev/mapper/vg_vm-lv_root exited with signal 11.
And now it's being rebooted unexpectedly showing following error:
SRAT: Hotplug area too small offline device
..MP-BIOS bug: 8254 timer not connected to IO-APIC
Kernel panic - not syncing: IO-APIC + timer doesn't work! Boot with apic=debug and send a report. Then try booting with the 'noapic' option.
I made snapshots, but rolling back to any of them does not help. As it turns out, this problem existed all the time, but it has become apparent only recently with more frequent rebooting and I/O errors.
I do not have official Red Hat support as it is the evaluation installation – I wanted to see how it works, test it and prove management I really need this OS. But all I want now is to save configuration of installed software in order to be able to move all data to another server. I cannot do this now – server is being rebooted 2 minutes after is has been started. No chance to export any data.
It seems that I'm stuck, I can't figure out what is the problem. Can someone explain what is happening? Maybe it was not configured properly on the HyperV virtual server? Can I fix it without OS reinstallation?
Best Answer
Error
SRAT: Hotplug area too smallis not related to your I/O problem, it's about memory in the SRAT hotadd area. I have such issues on all my Hyper-V guest systems working on CentOS with >= 3Gb of RAM. But I really don't know is it critical or not.