Hi, after my OS13.1 update crash I am thinking to install LEAP 42.1 on my PC.
I use a 250GB SSD as system disk, and 2 harddisk for the data. I harddisk ist for Win7 with NTFS and the other for OS13.1 with Ext4. Is there any known problem with this configuration and Leap 42.1 ?
I want to continue to use EXT4 and have read here how to make a SSD safe and fast. That’s from last year.
Does this apply also to Leap42.1 ?
I have 16GB of memory so no real need for a swap partition, right ? or wrong ?
On 02/15/2016 11:26 AM, J0EE0J wrote:
>
> Hi, after my OS13.1 update crash I am thinking to install LEAP 42.1 on
> my PC.
> I use a 250GB SSD as system disk, and 2 harddisk for the data. I
> harddisk ist for Win7 with NTFS and the other for OS13.1 with Ext4. Is
> there any known problem with this configuration and Leap 42.1 ?
None of which I am aware. NTFS support is pretty common, and has been for
years.
> I want to continue to use EXT4 and have read ‘here’
> (http://tinyurl.com/q7wp3ur) how to make a SSD safe and fast. That’s
> from last year.
Is there a particular reason you want ext4? As long as you’re doing
something fresh/clean/new, why not btrfs? The ext-based line of
filesystems, well ext3 primarily, have some really odd default settings
that make me want to kill a developer every time auto-fsck kicks in every
nth boot when I do not have time for it.
> Does this apply also to Leap42.1 ?
>
> I have 16GB of memory so no real need for a swap partition, right ? or
> wrong ?
Yes, I would probably ignore swap altogether unless you plan to use
hibernation of your box, in which case have as much swap as you have RAM.
–
Good luck.
If you find this post helpful and are logged into the web interface,
show your appreciation and click on the star below…
Hi, I’m running Leap on a 250 GB SSD (no spinning rust at the moment) and I see no problems so far…
That reference seems way too cautious IMHO, recent SSDs are not that fragile anymore…
I still have a 17GB swap for 16 GB RAM, you may still need it for performance in marginal cases (or for suspend-to-disk, not that SSD-killer if used from time to time); please note that if you run out of RAM, even temporarily, and you have no swap the system simply crashes.
To sum it up briefly:
EXT4 is OK
reducing swappiness is OK
FS trim is OK, might also be run on reboot if it is a laptop, say; cron is needed for servers that seldom reboot.
moving /tmp and /log to RAM is not necessary IMHO
noatime is OK
deadline scheduler is OK, but Leap should set it by default when it installs on a SSD (just check it)
browsers: I left them at default settings, wearout is no more a problem with normal use IMHO.
And there are several such threads in the forum, just search for them…
On 02/15/2016 11:56 AM, OrsoBruno wrote:
>
> I still have a 17GB swap for 16 GB RAM, you may still need it for
> performance in marginal cases (or for suspend-to-disk, not that
> SSD-killer if used from time to time); please note that if you run out
> of RAM, even temporarily, and you have no swap the system simply
> crashes.
No, that is simply not true. I run out of RAM once in a while when I try
to do something stupid, and the system is just fine. Applications crash,
notably the ones I’m trying to start that think they can have more RAM
than is available, but the kernel’s out-of-memory (OOM) kill does a good
job of finding the culprit even when a new application is not starting and
killing that one in particular (such as a memory hog).
–
Good luck.
If you find this post helpful and are logged into the web interface,
show your appreciation and click on the star below…
Technically correct, sorry for speed typing
APPLICATIONS are going to crash, not the whole system.
Nevertheless, even if not planning suspend-to-disk, I would still setup a (smaller?) swap to help the kernel with memory management and to have a warning from the system slowing down before actually beginning to kill processes, unless faced with special situations such as diskless thin-clients, very small disk or fixed applications like kiosks.
Once swappiness has been reduced, not having a swap space only because of possible SSD wearout is pure nonsense IMHO.
Anyway, these are just hints as the OP asked for: the actual setup is up to the system admin, thanks to the freedom Linux offers.
On Tue 16 Feb 2016 12:56:01 AM CST, OrsoBruno wrote:
ab;2754375 Wrote:
> No, that is simply not true. I run out of RAM once in a while when I
> try
> to do something stupid, and the system is just fine. Applications
> crash,
> notably the ones I’m trying to start that think they can have more RAM
> than is available, but the kernel’s out-of-memory (OOM) kill does a
> good job of finding the culprit even when a new application is not
> starting and
> killing that one in particular (such as a memory hog).
>
Technically correct, sorry for speed typing
APPLICATIONS are going to crash, not the whole system.
Nevertheless, even if not planning suspend-to-disk, I would still setup
a (smaller?) swap to help the kernel with memory management and to have
a warning from the system slowing down before actually beginning to kill
processes, unless faced with special situations such as diskless
thin-clients, very small disk or fixed applications like kiosks.
Once swappiness has been reduced, not having a swap space only because
of possible SSD wearout is pure nonsense IMHO.
Anyway, these are just hints as the OP asked for: the actual setup is up
to the system admin, thanks to the freedom Linux offers.
Hi
If push comes to shove, zram is always there… I have a systemd
service for it, but it’s just a script that creates ram swap space on
the fly…
–
Cheers Malcolm °¿° LFCS, SUSE Knowledge Partner (Linux Counter #276890)
SUSE Linux Enterprise Desktop 12 SP1|GNOME 3.10.4|3.12.51-60.25-default
If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below… Thanks!
