Samsung 860/870 SSDs Continue Causing Problems For Linux Users

[X_m7]

I own a couple of 860 SSDs. What am I supposed to do now, enable a periodic TRIM service (or something similar)?

This whole queued TRIM malarkey is for the continuous TRIM method, which is enabled if you see "discard" in /etc/fstab. Without queuing support, it can cause the system to freeze often, so yes you should switch to periodic TRIM ("sudo systemctl enable fstrim.timer" should do the trick). Apparently Ubuntu 18.04 onwards at least already use periodic TRIM by default, so you might not even need to worry about it.

Source: Glorious Arch Wiki

[puleglot]

Consumer platforms are unstable and not well validated in general, at least compared to server ones.

AMD SP5100 chipset (server variant of SB700) + Samsung 883 DCT (server variant of 860 EVO) have the same troubles with NCQ. This forced me to upgrade to a X570-based system several months ago.

[stormcrow]

This is a pity, as Samsung has traditionally been one of the better SSD makers, or at least the most recommended. I used to use Plextor's SSDs (though I suspect they are rebadges of someone elses drives), but I've used Samsung drives in a couple of builds so far and been happy with them.

You shouldn't be trusting Samsung more than any other SSD makers. They're pulling bait and switch with consumer devices just like others. This may or may not be related to the same problem Tom's Hardware & Extreme Tech have noticed, but I'm willing to bet it is. Crucial, Samsung, and others are known to switch out ICs after review release to cheaper parts. Who knows how long they've been doing this.

https://www.tomshardware.com/news/samsung-is-swapping-ssd-parts-too

https://www.extremetech.com/computing/326377-samsung-is-the-latest-ssd-manufacturer-cheating-its-customers

This may be SATA controller engineers expecting drive controllers to abide by the spec and the product release hardware doing no such thing, either by design or because of cheap substituted parts. Whatever the case, it bears a much closer investigation.

[drakonas777]

Yeah, you can always feel the confidence and 0 insecurities of the opponent, then they bring likes into the discussion LOL. I'm done with this. Pretty clear you have no interest neither in this specific issue nor in objectivity and keep pushing some general Intel-related ideology (or just trolling). Well, I'm not interested in that.

[perpetually high]

Maybe you should have empirical evidence that newer AMD chipsets suffer from this exact problem before doing such statements? In this case they should be marked for disabled NCQ as well. But they were not marked, which means there were no current hw bug reports. My statement is based on current data and not my personal experience (I use Intel BTW). Your statement, however, is based on intuition and fanboy shit "Intel CPU = no/less issues guaranteed".

If your intension is to troll, then sure, my mistake for responding somewhat seriously I guess.

Enjoy your likes from your buddies who care what you have to say. Not me, I don't. Said my piece. Intel has sat on no innovation for a long time, yes, but they are not the devil. Tired of that shitty narrative.

They (Intel) provide a lot for Linux (way more than shitty nvidia), and it's time to show some respect. Spectre/Meltdown was bad obviously, affected many not just them.

Unlike you guys, I don't blindly swear my allegiance to AMD. I highly, highly favor them though.

Carry on, boys. Would hate to go "off-topic" I know how much that bothers you guys to the core. Insecurities are keeping you guys disabled. Have a good one.

[perpetually high]

Sorry for going further off-topic, but I think Reddit's issue is mixing "like" and "dislike" and combining them into single number. In Youtube, they are separate and there's no such problem. (Well, Slashdot plays dice and doesn't let everybody vote. This is another method to prevent trolls messing things up.)

Don't be so scared, billy. No one's going to hurt you for going off topic. You have great ideas, you should post more.

That's a great point, thanks for bringing it up.

[thunderbird32]

This is a pity, as Samsung has traditionally been one of the better SSD makers, or at least the most recommended. I used to use Plextor's SSDs (though I suspect they are rebadges of someone elses drives), but I've used Samsung drives in a couple of builds so far and been happy with them.

What do you mean you don’t get the point of my statement? I don’t have a dog in the fight. Not everything has to be controversial, you know..

(edit: I truly feel like "Likes" becomes a popularity contest and a distraction to the conversation. Michael, you should really consider removing Likes and watch what it does to the discussion. If you have that capability, try it out. As an experiment. It provides nothing, except fake dopamine. Trust me, I know. I love all of your Likes.

and reddit is actually worse because fake accounts/trolls can quickly downvote a legimate comment and have it at -4 and then no ones takes it seriously, so they ignore it. Twitter doesn't have this problem. Shitty tweet? Zero likes, zero retweets. Please consider, Michael. Thanks for listening to my Ted talk.)

I occasionally look at the downvoted comments on Reddit, and 90% of the time they are: off topic, insulting, or complaining about downvotes on their previous downvoted comment. I'd rather a few controversial comments get buried than all that chaff rise to the top.

And speaking of off topic comments...

[mlau]

I feel like that's the computer version of WWJD.

True, but windows is probably the only OS the drive firmware gets extensively tested on, so knowing what windows does differently (it at all) is a win in my book.

[drakonas777]

As said by user drakonas777 about 5 minutes ago.

Maybe we should have some current APU, CPU, GPU, etc owners of AMD chips come forward and let you know about all of their problems. I know user sandy8925 mentions it all the time. I can go on further, let me know, but I don't want to keep hijacking threads but I won't be baited into a flame war when my motive and intentions were clear and have always been clear.

Maybe you should have empirical evidence that newer AMD chipsets suffer from this exact problem before doing such statements? In this case they should be marked for disabled NCQ as well. But they were not marked, which means there were no current hw bug reports. My statement is based on current data and not my personal experience (I use Intel BTW). Your statement, however, is based on intuition and fanboy shit "Intel CPU = no/less issues guaranteed".

If your intension is to troll, then sure, my mistake for responding somewhat seriously I guess.

[billyswong]

(edit: I truly feel like "Likes" becomes a popularity contest and a distraction to the conversation. Michael, you should really consider removing Likes and watch what it does to the discussion. If you have that capability, try it out. As an experiment. It provides nothing, except fake dopamine. Trust me, I know. I love all of your Likes.

and reddit is actually worse because fake accounts/trolls can quickly downvote a legimate comment and have it at -4 and then no ones takes it seriously, so they ignore it. Twitter doesn't have this problem. Shitty tweet? Zero likes, zero retweets. Please consider, Michael. Thanks for listening to my Ted talk.)

Sorry for going further off-topic, but I think Reddit's issue is mixing "like" and "dislike" and combining them into single number. In Youtube, they are separate and there's no such problem. (Well, Slashdot plays dice and doesn't let everybody vote. This is another method to prevent trolls messing things up.)