Skip to content
Register

Looking for host that doesn't cram 500 accounts per server

Reviews by Fritz48 7 replies 1K views
Fritz48

Member OP

Datenschutz First

14 #1

I am researching shared hosting density for a small GDPR-sensitive project. As far as I know, most providers do not disclose this figure. I have encountered:

  • One Hostinger sales rep who claimed "typically 80-120" but refused written confirmation
  • A RackNerd knowledge base article mentioning "dynamic density optimization" (meaningless)
  • InterServer's terms, which from memory have something about capping users per physical node at 400—but I think they define "node" as including multiple VMs I require actual numbers, not marketing architecture. My threat model includes co-tenant compromise via shared kernel or hypervisor escape. To be precise, I am willing to pay premium for verified isolation. Has anyone obtained real density figures under NDA, or measured indirectly via /proc or timing attacks?

Neuland. Aber schnell.
zoekyo

Member

Less Is More

2 #2

Complexity kills
If you have to ask the density, you already know the answer

rm -rf / --no-preserve-root ☯
moebig

Member

RAID is not backup

4 #3

Wesh Fritz48

Le server حلو, incha'allah for uptime, but the numbers ? They hide them like le trésor

I asked GreenCloudVPS once, they said "proprietary optimization" kkkkk. Proprietary means we dont want to tell you, véi

My trick : check the inode limit. 500k inodes = 500 accounts probably. 2M inodes = they cramming. Simple

RAID 1: because paranoia pays
hankels

Member

52 VPS and counting

3 #4

I got data

Measured indirectly: Hostinger "120 typical" is BS, I got 340 on a node once via /proc/loadavg patterns. CloudCone was worse, 600+ easy. KnownHost surprised me, actually hit their 200 claim in my spot-checks.

Best $/GB/RAM for actual low density? Leaseweb. Expensive tho. Like 3x. You pay for the air between accounts.

My list if anyone wants the chaos:

  • Hostinger: 340 real (claimed 120)
  • CloudCone: 600+ (claimed "flexible")
  • KnownHost: 190-210 (claimed 200) honest
  • Leaseweb: 80-90 (claimed 100) premium

seedbox, NAS, tape, and three offsite
olya

Member

Always Free

4 #5
hankels said:
Measured indirectly

Pro tip: your /proc method is detectable and violates most AUPs. Heads up before you get terminated.

Better approach for density research:

1. Check PTR records for neighbor domains (reverse IP lookup)
2. Count unique mail server banners on shared IP
3. Submit GDPR Article 15 request—some EU hosts actually respond with infrastructure specifics

I got actual VM counts from a German provider this way. Took 6 weeks, but the data was granular: 147 VMs on EPYC 7443, 512GB RAM, no overcommit.

For Oracle Cloud free tier: density is irrelevant, you get dedicated VM. The catch is the always-free vCPU throttling after month one. I wrote a guide on workaround patterns, DM if interested.

licensing is a suggestion
Sven3

Member

Stockholm

#6

Obviously hankels's "52 boxes" is a completely normal and healthy number of servers /s

I am wondering, or how do you say, whether olya's GDPR trick works for non-EU providers too. Clearly it would, because american companies love european bureaucracy. /s

Sven3 corrects Sven3: actually the sarcasm was obvious that time, no /s needed. Except I put it anyway. How do you say, self-awareness

debDebi

Member

#7

I Am Thinking The Oracle Is Right About The Vm Counts

I Am Running 12 Accounts On The HostHatch And I Am Thinking They Are Good With The Density. The Support Wrote Me "We Do Not Oversell" But I Am Thinking Every Host Says This.

I Am Checking The Inode Limit Like The Moe Said. The HostHatch Gave Me 4M Inodes. Is That Good Or Bad I Am Wondering

apt-get install everything
larryjeong

Member

#8

What's the inode limit on Hostinger

$3/year. 128MB RAM. Pure happiness.

Post a reply

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

Post reply Preview Save draft