A thing to watch for when mounting buckets locally

(experienced as a user not as a sno)
all was fine until the antivirus decided to scan all those vm backups.
and of course i had such a traffic that i perplexed.
yet another lesson learned.

seems obvious after explanation, not so obvious when you only see lots of traffic and have no clue why.

If on windows (not a windows user myself, just repeating old wisdoms), it apparently helps to use rclone mount --network-mode.

This is the problem. Uninstall it. You are wasting cpu and IO cycles on … what exactly? How many “viruses” have it caught for your latterly? And how much of a performance did you pay for it and continue paying with?

Antivirus is very useful as a revenue stream to antivirus companies. Most of which pivoted to “identity protection” business anyway.

Uninstall it. You don’t need it. Setup security and permission a properly. Antivirus won’t help you against social engineering and zero days. Relax.

if you would’ve been at least a millimeter competent, you would have known that windows defender is uninstallable.
and for the second millimeter, you would’ve known about folder exclusions.
and i am wasting more of the motoresource of my keyboard with you than cpu cycles with antivir.
you arrogant with me, me arrogant with you.

yea, first step in this is to convince the target to uninstall the antivir. :slight_smile:

Lol. Do you truly believe that? Are you sure? Excellent marketing, Microsoft.
(hint: turn it off in group policy)

I’m not saying exclude some folder. I’m saying uninstall the 100% useless bloatware. Checking folder exclusions also burns cycles. It is done in the filesystem filter. It’s sacrilege

Do you truly believe that keyboard polling is more resources than opening and scanning every file against pattern database and heuristics? Really?

Oh, so you truly believe it helps you? That’s the biggest part of the problem. It’s best not to have antivirus and know to be careful, than to have … whatever that is, and relax in false sense of security.

Just think about it.

Again, how many viruses did it catch for you in the last 10 years? And how much electicitity and hardware resource and your own time did you burn running it?

Diid you understand reference to zero days and social engineering? Antivirus does not help there. It does not help for most prevalent attack vectors, but you still waste cycles on it.

Am I? Does it matter? Does it change the applicability of what am I saying? Do you have technical arguments/rebuttal or you simply don’t like the tone?

even a couple months ago it caught some shit that pretended to be chrome updates, hehe.

oh that sense is always false, no matter antivir or not.
but antivir watches pretty well for the most obvious, of course.
just not the case with these huge vm backups about which it can alert me after a local scan and not the cloud-based one.

there is a power-off switch when i want a break from computers.
try that yourself too, it really helps a lot.

IIRC, antivirus software on Windows tend to ignore network drives because nobody wants it scanning potentially petabytes of remote storage from every single client that mounts it concurrently.

(yes, I’ve seen network drives with petabytes of data, we have one at work :sweat_smile: )

Oh wow.

If fake Chrome updates are reaching the point where antivirus has to catch them, you have a bigger issue you need to solve. Relying on the antivirus to protect you from yourself is not wise.

Antivirus shall not be the one that decides whether random downloaded executable is safe. You shall avoid that path entirely, fake browser update shall not have a way to get to you.

I’m not even going to ask where did you find fake browser update. I seriously don’t need to know.

(Separately, I would argue for not using chrome or any chromium derived product in the first place – but not for security reasons, this is a separate discussion. This is for the greater good for everyone – it repeats the IE saga, and we all know how it ended)

yea, but in my case it is mounted as a folder on an existing volume because rclone’s network drives are session-local and not system-wide!!!
most logical architecture of cloud-mounter i’ve tried is tntdrive but it does not support storj-native, only s3.

Sure, was just answering your question why would antivirus not want to scan it.

Also, I’m pretty sure I was doing an rclone mountpoint useful by a Windows system service long time ago. How, sorry, I don’t really remember—I’m trying to forget I have ever had to work professionally on Windows…

i know how but i personally dont want to go that route.
serivce wrappers.
that implies tedious work at even the slightest changes in executable parameters.
and it also means running rclone with local system privileges, a thing which should be minimized as possible and allowed only to very very signed executables. :slight_smile:

yea, everycrap in this world assumes you have it.
i was on os/2 when that happened with ie, but i know the story.

yea, and when that data should be accessed from several locations in the world, here comes storj.

and also there are those like me, apart from my backups, with my sudden spikes in storage needs when i experiment short-term with one thing or another so i prefer to rent and not to own.

Checked, yes, this option doesn’t work for folders and under system account it doesn’t allow to mount as a disk letter.

Good OS by the way, I used it as a client and as a server in the company where I was a system administrator, it worked much better than any available Windows that time..

i did an industrial automation project on it, apart from using it at home myself and for my hobby projects.
being a multithreaded os it helped me a lot and i avoided having very big state machines that you quickly lose control on.
very many threads (one thread per sensor plus several internal dataflow processors, spawned one thread per client data request connection) and a reasonable amount of mutexed access and event signaling among them.
and reliance that os gives access to mutexes based on thread priority; only this was on me instead of having my own priority queues and cracking my mind designing and managing them.
operator workplace was still on windows and made with delphi (gui charts involved), server-side that controlled the equipment was on os2 without workplace shell and written in virtual pascal (also an object pascal implementation), which had a brief current status in a console session that was configured to run full-screen.
all applicational software part was written by me, equipment was made by an electronics engineer.
yea, it was interesting to make it and the client used it many years after that.
it gave them unprecedented flexibility compared to the mechanical-only solution that they had before, so that they were very pleased.

as of those messages where i cursed and you asked me to edit, i couldnt edit (no such option anymore) so i deleted them.

as of pull requests, i dont think i qualify to try to mess with storj code, sorry.
there is still very little that i know about storj and cloud storage in general.
in some things i am a dinosaur, in others a noob!
and already getting old, having concentration problems.

I used REXX that time, because it’s a scripting language, but powerful enough to cover all needed automation for me.

Maybe you could try Object Mount for your task, it should be publicly available soon.

oh rexx was also involved, mainly for system maintenance tasks like logs rotation and stuff like that.

more of rexx usage was on my fido node at home, and at the previous job where i made the company’s site.
there it was apache with ssi in rexx, and it was presumed that my rexx-based prototype will be rewritten later in something like perl, but ibm released rexx for linux and that was installed on the production servers and my prototype was promoted into production without much blood. :slight_smile:

and the rexx substitution of today is called python, lol.

not much of a task on my side, OM would probably be more appealing to bigger fish than myself.

and i also shared my observations as a sno in a neghboring thread, maybe more nodes would consider switching to hashstore and i could benefit from that as a client and them benefit as sno same as i benefit myself as a sno.
hashstore really kicks ass performance-wise and looks like a win for all parties: clients benefit from improved speed and snos benefit from less hardware wear.