Score:0

A networked filesystem spontaneously mounted on my VPS and I'm not sure what, why or how

cn flag

This VPS was more or less idle during the days when this happened. The host (Scaleway) did not perform any maintenance or upgrades during this time, nor did I interact with the VPS directly or through the control panel. Automatic updates are disabled. The OS is Debian 10.

To me it looks like a networked filesystem was mounted for my VPS to access and enabled through some auto-discovery or similar, but I might be wrong. Is it perhaps the other way around, that the hypervisor/host somehow initiated my VPS to provide its storage over network for them to access? I understand they own the storage and can already access it in other ways.

This is what popped up in dmesg:

[3240105.606997] device-mapper: ioctl: 4.39.0-ioctl (2018-04-03) initialised: [email protected]
[3240107.026053] SGI XFS with ACLs, security attributes, realtime, no debug enabled
[3240107.066073] JFS: nTxBlock = 8192, nTxLock = 65536
[3240107.129470] QNX4 filesystem 0.2.3 registered.
[3240107.220733] raid6: sse2x1   gen()  5791 MB/s
[3240107.296739] raid6: sse2x1   xor()  4571 MB/s
[3240107.364749] raid6: sse2x2   gen()  9875 MB/s
[3240107.432744] raid6: sse2x2   xor()  5724 MB/s
[3240107.508740] raid6: sse2x4   gen() 11442 MB/s
[3240107.576734] raid6: sse2x4   xor()  6423 MB/s
[3240107.644738] raid6: avx2x1   gen() 11982 MB/s
[3240107.720722] raid6: avx2x1   xor()  6473 MB/s
[3240107.788733] raid6: avx2x2   gen() 13211 MB/s
[3240107.856730] raid6: avx2x2   xor()  7007 MB/s
[3240107.928766] raid6: avx2x4   gen() 14601 MB/s
[3240107.996734] raid6: avx2x4   xor()  7916 MB/s
[3240107.997462] raid6: using algorithm avx2x4 gen() 14601 MB/s
[3240107.998403] raid6: .... xor() 7916 MB/s, rmw enabled
[3240107.999125] raid6: using avx2x2 recovery algorithm
[3240108.008931] xor: automatically using best checksumming function   avx       
[3240108.111246] Btrfs loaded, crc32c=crc32c-intel
[3240108.129825] fuse init (API version 7.27)

Edit: title and premise is probably a misunderstanding on my behalf. What appears to be happening is that either the kernel for some reason suddenly loaded a few storage-related modules, or perhaps new hardware was connected to the host server while running.

Nikita Kipriyanov avatar
za flag
I don't see any mounting here. Various modules getting initialized, nothing else. And nothing network-related at all.
Speed Dial Dave avatar
cn flag
@NikitaKipriyanov OK, but how and why? I presume they manifested suddenly on the running kernel at 37 days of uptime for a reason.
Nikita Kipriyanov avatar
za flag
Is this some kind of a container-based system? Then *other* container (probably "root" system) might be doing something, but since all containers share the kernel, whole system has only one set of modules, when the module gets loaded, all containers suddenly realize it appeared. In properly set up system you simply can't see dmesg from non-privileged containers, because it's not your business.
Speed Dial Dave avatar
cn flag
@NikitaKipriyanov As far as I understand it's not container-based, but I cannot be sure. dmesg mentions: "efi: EFI v2.70 by EDK II", "DMI: Scaleway SCW-DEV1-MICRO, BIOS 0.0.0 02/06/2015", "Hypervisor detected: KVM" and other non-container-y things.
Nikita Kipriyanov avatar
za flag
You won't see that in dmesg. As I said, they share the kernel, so there is just one kernel log (which you see with dmesg), and all the containers see the same thing. You must just know. And VPS's almost always container-based which is **not** a virtualization, so there is no concept of hypervisor and the like. The whole question seems to me a very off topic, at least, in a way it is formulated.
Speed Dial Dave avatar
cn flag
@NikitaKipriyanov I can update and reboot to whichever kernel version I want. This is Debian 10 with all regular updates provided on "main" branch/repository. It's not locked to launch from some "boot script" with specific host-provided kernel. If this matters.
vidarlo avatar
ar flag
There's still nothing interesting in that log. There's no mounting going on. It's initialization of modules.
Nikita Kipriyanov avatar
za flag
If you own this system, you need to determine why some modules were initialized so late after boot. Usually this kind of initialization happens during boot; we can't know what was going in your system, but it's possible to trigger by just loading modules anytime using insmod or modprobe. In any case, the log you provided is **not related to mounting** and **not related to network**, there is nothing in common with the question title. If you want to have useful answers, please, ask the coherent question. Your hypotheses need to be at least in theory related to what you see.
Nikita Kipriyanov avatar
za flag
Again, this could be expected in VPS, because those are usually containers and you don't control what's going in root namespace of the system. If you're not the owner/operator of the *hardware*, dmesg is none of your business. It's wrong even that you are able to see it, and this automatically means any question about what you can see in it while only having an access to the VPS is incorrect and won't be answered here.
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.