New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Require Storage VPS in Singapore
Jovanfever
Member
in Requests
Looking forward for leads for storage VPS in Singapore region. Any providers?
Comments
https://hosthatch.com/products
Listen to the gentleman.
Top choice.
@remy Even though their website mentioned 7.5TB BW for 12 USD plan, the actual limit is 3TB per month for Singapore.
That is indeed true for now, however we're planning on making the bandwidth limits the same for all regions for our standard plans (before July 1st, and will apply to existing VMs as well).
Host hatch ontop
I can provide this for you in our SGP location with 10Gbit ports. Please let me know your RAM and HDD requirements.
You can test the network at https://sgp-lg.layer.ae/
Nice!!!
Even for old promotional plans?
@hosthatch
No unfortunately not. We already have higher bandwidth limits (compared to standard) for promotional plans.
Ok, just trying my luck
@hosthatch I'm looking for a storage VPS as well in Singapore for off-site backups. The intended use case is ZFS replication.
@remy @hosthatch Purchased one
The newer storage VMs are RAID 10. The older ones were RAID-50 or RAID-60 (just FYI).
Not as far as I know. It is a VM - so you can only "see" the RAID enabled volume as a disk within the VM.
Most of HH's nodes are 10+ Gbps - so you should see pretty solid uploads as well (subject to external network performance of course). I'm sure you can find many YABS all over the place for HH VMs (both storage and NVMe).
P.S: I've been a long term HH user with multiple storage VMs across locations and I've been quite happily using it for my ZFS needs (backups, replication etc.). Can't recommend them enough.
That's what I'm concerned about, since conventional raid increases the chance for bit rot.
Curious where you're located and which locations you have HH storage servers in?
How do you set up your zpool? My understanding is if the VM sees just one disk, you can only do a stripe. The only safe configuration I can think of is maybe setting copies=2.
That's why you (should) periodically scrub to ensure that what ZFS wrote is what it reads back again.
London, Amsterdam and Chicago.
With a single disk/volume. No redundancy (assumption is that the underlying RAID setup provides the redundancy and ZFS on top will do best-possible bit-rot detection, with no possibility to fix errors as it is a single "disk"). If the pool/dataset is a backup of another pool/dataset (somewhere else), at least you will know what is corrupt and can resync/resend and fix.
That makes sense. I'm new to ZFS - how does ZFS report bit rot to you? I assume you have an error message via the scrub, lets say it is in file abc. Does it tell you which snapshot of the file lives in, so you just zfs send that snapshot again to fix it? Is the zfs send smart enough to only send the bytes needed to fix tat specific problem block?
How often have you encountered scrub errors with HH?
you know zpool works on partitions and even files, right? you don't need a virtual disk.
the above creates a 200GB pool backed by a sparse file.
We have newer version nodes in LA, Chicago and New York now. All new storage VMs are set up on these. The new nodes come with RAID1/10 NVMe disks for OS, and a second RAID10 HDD disk for HDD.
We're going to do the same thing in Amsterdam, Stockholm, London, Singapore and Hong Kong, but it is going to take a few weeks/couple months. These locations still largely run RAID50/60, and some of them have RAID10 (but not NVMe OS disk), however regardless of the underlying RAID, please make sure you have at least two sets of your own data.
I forgot that was possible. So I can either have a pool backed by a single file vdev with copies=2, or a pool backed by a 2-file mirrored vdev. Is there a strong reason to choose one approach over the other?
@hosthatch are there any active promotions currently? I guess there isn't a convenient way for me to test upload speed from my location without having an actual instance first?
The answer is no to both of those questions unfortunately.
file-based vdevs allows you to control the parity. copies=2 or raidz1 with 2 vdev means you lose half the space. with a vps, all your virtual drives are backed by a single storage like RAID, so if you're going to lose data, you're going to lose all the drives, not a single virtual disk.
so the purpose of using zfs on vps is for bitrot, not virtual drive failures, and you don't need 100% parity for bitrot protection and lose half the space. You can create 5 file-based vdevs with raidz1 and you keep 80% of your space and leave 20% for bit-recovery.
That sounds great, I'll probably do this then. My only concern with this is write performance, which is something that I'd have to benchmark.