FragReady!
(EDITED)* So it seems that Fragready just knee jerks to bogus virus claims by fly by night idiots like  clean-mx.de because they are terrified of nethack on WindowsCE.  Yes really they deleted my server because of an old game on an old platform.
So here we go. again. Â 2013 seems to be the year of plenty of moves. Â While cruising around LEB, I came across this special on Frag Ready. Â So yeah I’m going to collapse all my VPS stuff (once it is finished copying) and move everything to a dedicated server.
What I’m hoping this will mean is that I can do far more neater things as now I don’t have to worry about CPU limitations, blowing my own quotas or being able to load whatever I want. Â I think I’ll even go back to offering some kind of public UNIX thing, I just have to decide if I want a SIMH VAX running BSD 4.3 UWisc, or whatever. Â I know I’ll certainly bring the Quake 1 server back, and maybe, just maybe hack enough to get a Doom dialup server going (if I can convince it to talk to my fake modems).
Another observation is that using the new ext4 filesystem means things are slower than ever. Â I know this server is two years old but still my seven year old Mac Pro destroyes it running Qemu vs running KVM on this linux box. Â I’ve found the two things help for performance some.
Convert disk images from sparse VMDK to QCOW2
|
# qemu-img convert -f vmdk -O qcow2 -o preallocation=metadata source.vmdk destination.qcow2
|
And changing KVM from ‘-hda disk.vmdk’ to
|
kvm -cpu pentium -m 256 -drive file=/usr/local/kvm/disk.qcow2,if=ide,index=0,media=disk,cache=none -vnc :0 -net nic,model=pcnet -net user
|
Next was to change the way the volume was mounted. Â First a change in the filesystem
|
tune2fs -o journal_data_writeback /dev/sda1
|
Then changing the options to the following in fstab:
|
noatime,data=writeback,barrier=0,nobh,errors=remount-ro
|
So yeah, it feels a little better now.
Here we go, again with what is moved over so far:
- BBS Migrated
- Old Linux Migrated
- Unix Migrated
- UTZoo in Migrated
- efylon.org Migrated
- vpsland Migrated
- macmint Migrated
- jdosbox Migrated
- zippy Migrated
- virtuallyfun.com Migraged
- Blog Migrated
- QuakeWorld Back to Life!
If you need to change again, friend suggests RamNode.
And ext2 is faster than ext4 – journaling, remember. There’s other FSes like ZFS, XFS, btrfs, etc.
Also, public BSD!
I was driven by price and core count… 8 cores is nice. As soon as I can justify I’ll get KVM access and use lvm for raw partition access. Or just install proxmox ve
xenix vm maybe? yeah, i’m a troll sometimes. : )
it certainly is a possibility…. depending how reliable that SLIP line can be….
If I had the disks for it…I’d bring up 4.3BSD on a MicroVAX II. Also the power.
There is always simh…..
That’s not as fun as the real thing though. 😉
Yeah… I still hope one day to get an 11/780 to run BSD on. But the space/power for that is utterly insane!
Very true. I’ve run in to people with 11/780s, 11/750s, and 11/730s, however.
probably all running VMS, on Hecnet.. lol
Yup. I need to have them help me get VMS 1.5 rolling on simh.
ZFS, ftw. Besides that: I like what you did with this place.
Cheers,
-joe
I tried zfs on linux once.. It’s a userspace module… And very slow. Also after a few GB of traffic and reboots it just died on me. I’d be scared of zfs on solaris… Notice how apple backed way away from zfs.
Otherwise, yeah the server is nice.
It’s been turned in to a kernelspace driver now. ZFS’ point is to be more reliable than other filesystems. Solaris is where it started so it works fine THERE.
Lost data with Solaris 10 + ZFS. Shame on me. Lost data with Linux + ZFS userspace, shame on me twice. Not going to try ZFS ever again.
Besides its a single disk, utterly pointless with ZFS
Yeah, single disk does become fairly pointless.
For a while I was really happy with this setup. But then they deleted my server because of some bogus ‘virus scanner’. Oh well.