Yeah, but a) I'd have to drive across town to get my rope and b) Last time it was windy and I had it on it flapped about like a sail and blew the whole thing over
So the unwanted DDoS they're insistent on forcing on everyone (because they seem to attract a lot of uber popular ddos targets as customers) is holding things up. Hmm.
#
# DO NOT EDIT THIS FILE
#
# It is automatically generated by grub-mkconfig using templates
# from /etc/grub.d and settings from /etc/default/grub
#
### BEGIN /etc/grub.d/00_header ###
if [ -s $prefix/grubenv ]; then
set have_grubenv=true
load_env
fi
if [ "${next_entry}" ] ; then
set default="${next_entry}"
set next_entry=
save_env next_entry
set boot_once=true
else
set default="0"
fi
if [ x"${feature_menuentry_id}" = xy ]; then
menuentry_id_option="--id"
else
menuentry_id_option=""
I actually never even touched grub.cfg on my MG-128
notice how the identifiers in the search commands were automatically changed, by update-grub, to 5ba39976d1877d83?
those used to be that c8... identifier that I get an error about when trying to boot
c823457... rather
oddly, neither blkid nor /dev/disk/by-id nor /dev/disk/by-uuid contain any substring with 5ba399 in it
First, the bios_grub flag in parted or GParted is a way to identify a BIOS Boot Partition. GRUB 2 uses this partition to hold part of its boot code when the computer boots in BIOS mode. It's normally about 1MiB in size, although it can be smaller in some cases. Most importantly, if Boot Repair re...
When a GPT partition table is present on the system, a small BIOS boot partition with type EF02 (which is different from the EFI System Partition (ESP) which has type EF00) will need to be available. 1 MiB will be enough to work, but 2-4 MiB is a safer option. This BIOS boot partition will hold the stage 2 of the bootloader. BIOS boot partitions do not need to be formatted with a filesystem; the grub-install command will overwrite any existing filesystem with one of its own.
I think the only problem with this method (BIOS boot with GPT) is that after the disks reach the ~4TB size, they won't be bootable in CSM mode any longer.
I'll eventually (once OVH offers bigger disks as standard) have to figure out a way to go UEFI boot -> grub2 -> ZFS root
well if they want to give me a free upgrade to 4 TB I definitely wouldn't turn them down... but I'm not super interested in having to redo my server now that I just got it working perfectly
I have 990 GB free space and I can definitely delete some stuff and clear out like 200 GB with some effort
I'm already reaping the benefits of a 25% off coupon code on this new server
and downgrading from the MG-128 Haswell to the SP-128-S Ivy Bridge, plus the 25% off, has saved me about $1750 per year
that's half a laptop
nice low CPU load even with all services running
hmm, what should I use my MG-128 for for 8 days? seems sad to let it go to waste
One thing the DNG converter does that I'm not particularly fond of is the fact that it appends _1 to filenames when they already exist. Given the number of files I'm dealing with, this is a huge pain. To work around this, I've decided to set it to output to another drive so that I can just overwrite data when moving files back.
(1) Yes, always "consider" it; whether to actually do it or not is up to you. (2) You can get a pretty decent dedi at OVH within the $50-ish/month price range - not bad. (3) I use it for Cavil, for hosting a website for someone else, for general file storage, as a proxy/VPN behind restrictive networks, for getting access to a Windows box (VM) from my phone, for hosting games (on and off, not 24/7/365), for my custom Cards Against Humanity fork, for music streaming to friends once or twice a month, and for development and learning.
@bwDraco you can play with ZFS using a tiny instance; you don't need any special hardware to play with ZFS unless you want multiple storage devices, but even then you can do a "filesystem in a file"
the amount of hardware you get for what you're paying is about 10% of what you could get at OVH or Hetzner
even AWS is better, though not by too much, as it's still a VPS
VPSes are basically an absolute crapshoot unless (a) you're too lazy to manage containers / VMs on your own hardware, or (b) your needs are so ridiculously modest that you can get by on a cheap $10/month plan; you can't buy a dedi that cheap
if your budget is $50/month or higher I can't ever recommend getting a VPS unless you prefer to sacrifice WAY worse hardware resources for potentially less administrative overhead (and less likely to do something like brick networking requiring you to use IPMI to go fix it)
the benefit really isn't that much these days with LXD and KVM (my recommended containers and VM solutions, respectively) being so good
the closest is Beauharnois, Canada, which is in west-southwestern Montreal... you can get there by car in 6 hours from Louis Rossmann's repair shop in downtown NYC, according to Google Maps
C:\Users\Brian>ping -t bhs.proof.ovh.net
Pinging bhs.proof.ovh.net [2607:5300:60:44a5::1] with 32 bytes of data:
Reply from 2607:5300:60:44a5::1: time=43ms
Reply from 2607:5300:60:44a5::1: time=44ms
Reply from 2607:5300:60:44a5::1: time=44ms
Reply from 2607:5300:60:44a5::1: time=42ms
Reply from 2607:5300:60:44a5::1: time=48ms
Reply from 2607:5300:60:44a5::1: time=38ms
Reply from 2607:5300:60:44a5::1: time=46ms
Reply from 2607:5300:60:44a5::1: time=47ms
Reply from 2607:5300:60:44a5::1: time=44ms
That's kinda okay, but this is what I'm getting right now:
C:\Users\Brian>ping -t www.fierydragonlord.com
Pinging www.fierydragonlord.com [2600:3c03::f03c:91ff:fe69:609a] with 32 bytes of data:
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=26ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=27ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=20ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=16ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=24ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=20ms
Reply from 2600:3c03::f03c:91ff:fe69:609a: time=20ms
I haven't seen a huge need for more compute power right now, so I'll stick with Linode, but will definitely consider dedicated servers for future needs.
I'm not in the mood for server migration.
Experience with administering a server on bare metal and virtualization technologies would be great for a future job, though.
It looks like currently most OS installers insist on putting /boot on a non-RAID partition (or the kind of RAID1 partition that "looks like" a non-RAID partition), even the installers that support RAID5 and GRUB2.
I'm guessing this limitation is a historical relic leftover from Grub1.
My underst...
That's probably going to be a huge learning curve, with no applications in mind for the server right now.
Perhaps we'll just treat this as a toy project? My main Web server is likely to stay on Linode ($10/month), but my cloud desktop is likely to go to the dedicated server once properly set up.
$56/month, or even $40/month, is a lot to pay for a toy.
Then again, it's much more than a toy; I've used the experimental cloud desktop for things like (semi-)secure remote printing and productivity in the past.
PSA: Please do not use flags for non-offensive content simply because you disagree with it. Everyone is entitled to their opinion and flags are only for genuinely spammy or offensive content. The misuse of flags is disruptive to high-reputation users on all of the chatrooms on this server.
I don't have the space furniture to set up my old eMachines desktop right now.
I have an extra 250 GB SATA hard drive lying around I could use...
I can give the VM up to 16 GB of memory although that might starve the host machine a bit (there's 24 GB on this laptop).
I might create a bunch of different virtual disks, with ZIL and L2ARC as one virtual disk on the system's main SSD and the main zpool as two virtual disks on the hard drive.
Any printable PDF documents on how to get started with ZFS?
Maybe if you ran enterprise level stuff on the ZFS system, but merely for testing? Any consumer SSD will do.
Ok, maybe not any. There are some truly broken SSDs out there. But most are either first generation SSDs or TLC.
And sane SSD (enterprise or simply std consumer) has plenty lifetime in it. You might need to replace them in 5 to 15 years.
But by that time most are obsolete anyway.
As to power loss protection: If you want to be safe: aye. For testing, meh, just do not thow the power switch and expect things to be safe
And before it is asked, no ECC needed either (though SLC, ECC and prosumer stuff would be nice to play with. It is just not needed)
The only thing you want for ZFS is memory. ANd more memory. And not run FreeBSD with root on FFS and the rest on ZFS. And no HW RAID and not turn on de-duplication
I have a folder where subdirectories and files exist. I want to ensure no two directories have same name. How can do this?
