Skip to content

Monitoring disk health on KVM VPS without SMART access

VPS Hosting by frankfurt_ops 16 replies 2.3K views
#1

The Problem: Virtual Machine Disk Images hide Physical Media Health from the Guest. SMART Tools return no Data. The Provider will not share Backplane Status.

My Workaround for Time4VPS and similar Providers:

  • Filesystem Level: enable Data Checksums (btrfs or zfs if available)
  • Application Level: verify Checksums on stored Archives monthly
  • Service Level: monitor Read Latency Spikes with iostat; sudden Degradation predicts underlying Array Issues
  • Provider Level: parse Status Pages and API Endpoints for Maintenance Windows

Please to run this monthly:

find /data -type f -exec md5sum {} + > checksums.new && diff checksums.old checksums.new

This is not perfect but it is what we have. Please to share Improvements.

Hans-Georg Mueller
2026-01-19 09:14 CET

#2
frankfurt_ops said:
Please to share Improvements

Back when we ran bare metal, SMART was gospel. Kids these days rent black boxes and pray. 20 years in this business, I have seen two providers silently migrate degraded arrays and only announce it after customer data evaporated. Your checksum approach is sensible but insufficient. Mark my words: the next evolution will be providers offering blockchain-verified integrity proofs at extra cost, and everyone will pay because the alternative is blind trust. The real fix is owning your physical layer, but that ship sailed for this market segment.

IPv4, IRC, and irssi — fight me
#3

Asking for a friend who wants that physical layer

Seriously though, Leaseweb has an API endpoint for storage health, undocumented but I found it

curl https://api.leaseweb.com/v1/disk/status

Returns json with synthetic health score. Not SMART but better than nothing. Privacy coins accepted btw

not your keys, not your coins
#4

Thread locked. Disk health discussion belongs in dedicated monitoring subforum. Take provider API scraping to DMs before terms of service violations become bans.

#5

What does Time4VPS charge for that checksum tier

phở at 3AM, deploy at 4
#6

Two providers or two arrays

airgapped, encrypted, faraday'd, still worried
#7
olespete said:
Two providers or two arrays

Two providers. One in 2011, one in 2019. First was a "premium" cPanel host that shall not be named, second was a European brand that got acquired and asset-stripped. Both times the migration happened during a "routine maintenance" window, both times the degraded source array threw unrecoverable read errors that the hypervisor passed straight through as silent corruption. My checksums caught it on restore, not on live data. That's the gap frankfurt_ops's monthly diff won't close: you only know after the damage is done.

IPv4, IRC, and irssi — fight me
#8

Time4VPS Storage VPS starts at 2.99 EUR, standard KVM at 3.49 EUR. No separate checksum tier, you just get the disk they give you. I run three in Vilnius for a client who wanted EU data residency. Latency to HCMC is 280ms, not great, but the price is the price.

phở at 3AM, deploy at 4
#9
haroldgsm said:
You only know after the damage is done

This is why I run the diff weekly now, not monthly, and keep three rotating checksum files on a separate provider. The real gap is detection speed, not method. If anyone wants to split a Hetzner Storage Box for offsite checksum mirrors, DM me.

#10
frankfurt_ops said:
Keep three rotating checksum files on a separate provider

Paranoid. Correct. I do the same with rsync to a BuyVM slab in Luxembourg. Different ASN, different power grid, different everything. The checksum file is 4MB for 800GB of data, trivial to move.

airgapped, encrypted, faraday'd, still worried

Post a reply

You need an account to reply. Log in or register to join the conversation.

Post reply Preview Save draft