Linux - GeneralThis Linux forum is for general Linux questions and discussion.
If it is Linux Related and doesn't seem to fit in any other forum then this is the place.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
View Poll Results: What filesystem do you use?
Extn
60
84.51%
Reiserfs
1
1.41%
Xfs
9
12.68%
Jfs
5
7.04%
Btrfs
15
21.13%
Other
5
7.04%
Multiple Choice Poll. Voters: 71. You may not vote on this poll
I'm assuming that SSDs don't actually have lifespans measured in days and that Linux works better when there are file permissions. I'm also assuming that on the majority of systems a journal, as opposed to bear FAT, is at worst useless during an unexpected failure and at best a little help.
I'm assuming that dave@burn-it.co.uk will come up with a niche or edge case and a story about SSDs failing faster then they should.
I've never read about issues with journaling file systems causing data loss either, nosiree bob. I know them to be as 100% reliable as FAT.
As long as you make sure to use noatime option on SSD's, their lifespan is quite decent, and will probably outlast other components with normal usage. Newer SSD's supporting TRIM increases their theoretical lifespan. Obviously, there will be cases with SSD's dying quicker, as with everything that is inherent on hardware that cannot possibly have a 100% quality rate.
Journalling file system were developed back in the day when disks and more importantly the transmission of data were unreliable and retries were the norm rather than the exception.
Nowadays they are only required in situations where there is a high likelihood of disruption to the media or data lines, though with the increasing trend towards distributed processing I suppose the requirement is again increasing.
Journalling is bad news for solid state drives especially with small buffers or highly random data because of the number of writes generated, not just to the individual cells but to the whole blocks which must be rewritten each time. It also slows down data transfers. Don't forget that solid state cells cannot just be overwritten, they have to be set to zero and then the new data written and that will be the WHOLE BLOCK even if only a single bit is involved - which won't be as bad if the block just happens to be in the buffer.
Writing a journal is more than doubling the likelihood that blocks will need to be rewritten and is doubling the likelihood that an error is going to occur in the first place. Modern CRC checking of data pretty much removes the need for it except where errors are more likely during transmission and even then error should be trapped.
Last edited by dave@burn-it.co.uk; 06-21-2016 at 02:37 PM.
I'm a JFS user on pretty much all my boxen; desktop and laptop, on SSD drives, thumbdrives, and everything else. I'm quite pleased with it, its inability to be resized notwithstanding. It has proven to be stable, graceful in recovery, and good enough for the multimedia applications I use. No complaints, no reason to move away from it.
On SSD, I do set noatime and I utilise the latest JFS with its TRIM optimisations.
My oldest SSD-based system currently has a total power-on time of about 2300 days and it's still going just fine, chugging along 24/7.
I have a 64GB SSD. smartctl reports mine has 40396 hours on it. Runs nearly 24/7. Reports only 8% of it's life used. It's had luks/ext4 the whole time. A journaled filesystem shouldn't be an issue for modern SSDs.
Distribution: Debian Sid AMD64, Raspbian Wheezy, various VMs
Posts: 7,680
Rep:
Quote:
Originally Posted by dave@burn-it.co.uk
Journalling file system were developed back in the day when disks and more importantly the transmission of data were unreliable and retries were the norm rather than the exception.
Nowadays they are only required in situations where there is a high likelihood of disruption to the media or data lines, though with the increasing trend towards distributed processing I suppose the requirement is again increasing.
Journalling is bad news for solid state drives especially with small buffers or highly random data because of the number of writes generated, not just to the individual cells but to the whole blocks which must be rewritten each time. It also slows down data transfers. Don't forget that solid state cells cannot just be overwritten, they have to be set to zero and then the new data written and that will be the WHOLE BLOCK even if only a single bit is involved - which won't be as bad if the block just happens to be in the buffer.
Writing a journal is more than doubling the likelihood that blocks will need to be rewritten and is doubling the likelihood that an error is going to occur in the first place. Modern CRC checking of data pretty much removes the need for it except where errors are more likely during transmission and even then error should be trapped.
So, you have machines which never undergo unclean shutdowns, have old SSDs which don't last very long, don't beneit at all from having file permissions and never need to use files larger than 4GB.
Fair enough, unusual, but fair enough.
So, you have machines which never undergo unclean shutdowns, have old SSDs which don't last very long, don't beneit at all from having file permissions and never need to use files larger than 4GB.
Fair enough, unusual, but fair enough.
My thoughts exactly...
But then that doesn't qualify as actually using GNU/Linux, only pretending to do so (flame-suit buttoned up)...
I have a 64GB SSD. smartctl reports mine has 40396 hours on it. Runs nearly 24/7. Reports only 8% of it's life used. It's had luks/ext4 the whole time. A journaled filesystem shouldn't be an issue for modern SSDs.
Pretty much the same here with my home server. 40 GB Intel SSD running the OS and everything but /home (that's on a RAID). 44800 power-on hours, 3% of its life used, running ext4 with noatime.
The wear-limit paranoia surrounding SSDs 5-6 years ago has been thoroughly debunked in recent years. Under regular usage they'll last as long or longer than any other part in the computer, barring random failures that can occur in any device. You'd have to try, hard, to actually wear out an SSD during the lifetime of the rest of the computer without doing benchmarks or dd dumps 24/7. Filesystem journaling won't come close to doing it.
Last edited by suicidaleggroll; 06-21-2016 at 04:20 PM.
But then that doesn't qualify as actually using GNU/Linux, only pretending to do so (flame-suit buttoned up)...
You're right to have that flame suit. What happened to the idea of freedom of choice? Is the GNU/Linux philosophy that it allows you to have freedom of choice, but only if you do everything its way?
Bravo, I say, to everyone who bucks the trend and uses the tools that GNU/Linux supplies in unusual ways. That's the way we discover things.
You're right to have that flame suit. What happened to the idea of freedom of choice? Is the GNU/Linux philosophy that it allows you to have freedom of choice, but only if you do everything its way?
I did not suggest that they were not free to do it that way, I simply suggested that not using something does not constitute using it... it is pretending.
Everyone is free to use or abuse it as they see fit. But if I put Ford seats in my Chevy SUV and then claim to be a Ford driver, and offer others advice on using their own Fords, I am then abusing other people's intelligence more than anything else.
Quote:
Originally Posted by hydrurga
Bravo, I say, to everyone who bucks the trend and uses the tools that GNU/Linux supplies in unusual ways. That's the way we discover things.
My point exactly, in a twisted logic kind of way!
I say again, not using the tools that GNU/Linux supplies, does not constitute using the tools that GNU/Linux supplies!
Not using something does not, in general, help you discover how it works!
I did not suggest that they were not free to do it that way, I simply suggested that not using something does not constitute using it... it is pretending.
Everyone is free to use or abuse it as they see fit. But if I put Ford seats in my Chevy SUV and then claim to be a Ford driver, and offer others advice on using their own Fords, I am then abusing other people's intelligence more than anything else.
My point exactly, in a twisted logic kind of way!
I say again, not using the tools that GNU/Linux supplies, does not constitute using the tools that GNU/Linux supplies!
Not using something does not, in general, help you discover how it works!
Ah, but the poster in question *is* using Linux, just not quite to the fullness of its abilities (but then again who does?).
Do you know, until the poster mentioned that they installed Linux on a FAT drive, I had no idea whatsoever that it was even possible (given for one thing the fact that FAT doesn't support the file attributes that POSIX-like systems do). It was an eye-opener. Good for them.
Ah, but the poster in question *is* using Linux, just not quite to the fullness of its abilities (but then again who does?).
Do you know, until the poster mentioned that they installed Linux on a FAT drive, I had no idea whatsoever that it was even possible (given for one thing the fact that FAT doesn't support the file attributes that POSIX-like systems do). It was an eye-opener. Good for them.
Well, they CLAIM to be using linux. Every post (in this thread) was from a Win7 system (if the user agent of their browser is accurate), no linux.
Well, they CLAIM to be using linux. Every post (in this thread) was from a Win7 system (if the user agent of their browser is accurate), no linux.
:-D I don't think they were making all that up.
We live in a complex computing world with devices running all manner of operating systems. I didn't realise that you weren't taken as seriously if you posted in Linux Questions from a non-Linux device (or didn't make sure that your browser agent was reporting the fact correctly). I must watch out for that.
We live in a complex computing world with devices running all manner of operating systems. I didn't realise that you weren't taken as seriously if you posted in Linux Questions from a non-Linux device (or didn't make sure that your browser agent was reporting the fact correctly). I must watch out for that.
You're reading FAR more into my post than I meant. I'm just saying, there's quite a few users that post here that DON'T use linux. Most of them are trolls that only stay for a few months to badmouth linux then move on. Not saying he's one of them, just saying his useragent shows he's always using Windows. He may have ran linux and decided he doesn't like it. He may still run linux but only for certain things and none of them need browsers. It's just that when someone makes 5+ posts in a linux thread over several days mostly talking bad about some of the core technologies, and never shows using linux on any of them, I personally feel less than overwhelmed to believe that person actually uses linux.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.