Swap Partition Vs Swap file

Hi everyone!!
I am using Oracle 11g (enterprise edition) on linux redhat 5.5 . At present, we have 12G RAM and swap partition of 2G. I know that we should have swap area of 6GB as per RAM. But we are unable to create new swap partition or resize existing. So I want to know is should I create new swap file of 4Gb that can be used as swap area in addition to existing swap partition of 2Gb. So that total swap area will be of 6G
( 2G swap space + 4G swap file).
1. Will it be meaningful or not ?
2. Which one is better among swap partition and swap file ?
3. Will swap file cause performance degradation due to incompatibility between them ?
Thanks & Regards
Tushar Lapani

Swap is no substitute for RAM. It provides a fail-safe mechanism and allows the kernel to handle active memory allocation more efficiently.
Your requirement or conclusion about swap space may or may not be reasonable, depending on your desired configuration. As a rule of thumb, swap space is usually configured twice the RAM if you use 8 GB or less and equal the amount of RAM if you have 16 GB or more. If you have 64 GB of RAM, it does however not make sense to have 64 GB of swap space, because the performance degration should you ever need that much swap space will be very bad.
A swap file as opposed to a swap partition has some security and safety concerns, and some performance overhead because of using a file system, but it can be feasible. It depends on your I/O use and hardware.
However, you will probably want to configure your Oracle database to use kernel Hugepages using ASMM instead of AMM for best performance and efficient memory use. Kernel Hugepages, unlike POSIX shared memory (/dev/shm) used by AMM, will be pre-allocated at system startup and cannot use swap space.
Message was edited by: Dude!
Until kernel 2.4 it was considered that a swap partition is faster than a file, but with the improvements of kernel 2.6, the performances is almost the same. A swap file has the advantage that it can be easier re-sized or removed than a swap partition, but it can get fragmented or deleted by accident. For best performance it should also be located at the outer parts (beginning) of the disk.

Similar Messages

Swap partition VS Swap File system

Hi everyone!!
I am using Oracle 11g (enterprise edition) on linux redhat 5.5 . At present, we have 12G RAM and swap partition of 2G. I know that we should have swap area of 6GB as per RAM. But we are unable to create new swap partition or resize existing. So I want to know is should I create new swap file of 4Gb that can be used as swap area in addition to existing swap partition of 2Gb. So that total swap area will be of 6G
( 2G swap space + 4G swap file).
1. Will it be meaningful or not ?
2. Which one is better among swap partition and swap file ?
3. Will swap file cause performance degradation due to incompatibility between them ?
Thanks & Regards
Tushar Lapani

This is a 100 percent Unix question and has nothing to with Oracle RDBMS.
Please close this question here and post it in any Unix forum.
Sybrand Bakker
Senior Oracle DBA

[Solved] Do I need a swap partition /swap file?

Hello Arch Community.
I want to install Arch Linux on my Aspire V5-573G Notebook with 8GB RAM and a 120GB 840 Evo SSD from Samsung. So my question is, do I actually need a swap partition / swap file? I read the arch-wiki pages about swap and suspend-to-disk / suspend-to-ram, and the only thing thats obviously profiting from / needing a swap partition is suspend-to-disk, no other things are mentioned in this regard. Is a swap partition in any other way important?
Thanks in advance.
Last edited by Oedner (2013-12-21 12:41:19)

headkase wrote:Another question to ask is what will happen if you don't have a swap file and run out of memory? In that case the Linux kernel will randomly kill process' until there is enough memory again. Obviously that isn't very stable but that is what it will do.
No. The OOM killer is pretty darn tweaked these days and usually(!) kills the actual perpetrator. The alternative—disabling memory overcommitting and answering memory requests with an out-of-memory error—can result in unforeseen problems as well: What if it a system process is trying to allocate a few bytes for some random string, and what if said system process isn't equipped to handle ENOMEM gracefully and crashes, because the default policy on Linux is to overcommit and never answer with out-of-memory errors?
Just trying to play devil's advocate here. I'm not happy with the OOM killer either and I'd rather have a deterministic solution to out-of-memory situations, but I realize that it's probably not an easy problem to solve.

[SOLVED] Virtual Memory Exhausted with unused 16g Swap Partition.

Hello all, long time reader, first time poster. So basically what I'm running into is that I'm having Virtual Memory Exhausted errors with various programs even though I have a 16g swap partition mounted. I've manually enabled it to try it as well. Here's some outputs to give you an idea of how my system is set up. It's a UEFI fresh install, only a couple days old.
fstab:
# /etc/fstab: static file system information
# <file system> <dir> <type> <options> <dump> <pass>
tmpfs /tmp tmpfs nodev,nosuid 0 0
# /dev/sdb3
UUID=ea7a0311-6314-4215-b98d-7360552b28c4 / ext4 rw,relatime,data=ordered 0 1
# /dev/sdc3
UUID=8f6f2006-3a9c-4563-a149-547d758e2af1 /home ext4 rw,relatime,data=ordered 0 2
# /dev/sdb2
UUID=baf1ba80-ddf3-4cbb-a13e-b81a0af0a9e7 /boot ext4 rw,relatime,data=ordered 0 2
# /dev/sdb1
UUID=2B46-5340 /boot/efi vfat rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro 0 2
# /dev/sdc1
UUID=4336d783-c289-4fbf-8c59-f1a1da782130 /var ext4 rw,relatime,data=ordered 0 2
# /dev/sdc2
UUID=96a60f08-a57f-42f1-80ff-846d02a29199 /var/log ext4 rw,relatime,data=ordered 0 2
# /dev/sda2
UUID=c5d5879a-b0e8-4976-8057-2f95975a1251 /share ext4 rw,relatime,data=ordered 0 2
# /dev/sdb4
UUID=6e6277c4-a69a-4838-83cd-c206959acf0e /store ext4 rw,relatime,data=ordered 0 2
#swap
UUID=21a56b10-85e5-48a1-91e5-995eae0dfba4 none swap defaults 0 0
lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 931.5G 0 disk
├─sda1 8:1 0 16G 0 part [SWAP]
└─sda2 8:2 0 915.5G 0 part /share
sdb 8:16 0 931.5G 0 disk
├─sdb1 8:17 0 1G 0 part /boot/efi
├─sdb2 8:18 0 1G 0 part /boot
├─sdb3 8:19 0 50G 0 part /
└─sdb4 8:20 0 879.5G 0 part /store
sdc 8:32 0 931.5G 0 disk
├─sdc1 8:33 0 20G 0 part /var
├─sdc2 8:34 0 2G 0 part /var/log
└─sdc3 8:35 0 909.5G 0 part /home
sr0 11:0 1 1024M 0 rom
cat /proc/meminfo
MemTotal: 12256600 kB
MemFree: 9926584 kB
Buffers: 212588 kB
Cached: 774636 kB
SwapCached: 0 kB
Active: 1179860 kB
Inactive: 638584 kB
Active(anon): 831604 kB
Inactive(anon): 14480 kB
Active(file): 348256 kB
Inactive(file): 624104 kB
Unevictable: 4 kB
Mlocked: 4 kB
SwapTotal: 16777212 kB
SwapFree: 16777212 kB
Dirty: 32 kB
Writeback: 0 kB
AnonPages: 831320 kB
Mapped: 196008 kB
Shmem: 14892 kB
Slab: 345008 kB
SReclaimable: 316620 kB
SUnreclaim: 28388 kB
KernelStack: 2616 kB
PageTables: 27032 kB
NFS_Unstable: 0 kB
Bounce: 0 kB
WritebackTmp: 0 kB
CommitLimit: 22905512 kB
Committed_AS: 3365808 kB
VmallocTotal: 34359738367 kB
VmallocUsed: 357936 kB
VmallocChunk: 34359373819 kB
HardwareCorrupted: 0 kB
AnonHugePages: 86016 kB
HugePages_Total: 0
HugePages_Free: 0
HugePages_Rsvd: 0
HugePages_Surp: 0
Hugepagesize: 2048 kB
DirectMap4k: 117756 kB
DirectMap2M: 4059136 kB
DirectMap1G: 8388608 kB
swapon -s
Filename Type Size Used Priority
/dev/sda1 partition 16777212 0 -1
core file size (blocks, -c) 0
data seg size (kbytes, -d) unlimited
scheduling priority (-e) 20
file size (blocks, -f) unlimited
pending signals (-i) 94832
max locked memory (kbytes, -l) 64
max memory size (kbytes, -m) unlimited
open files (-n) 1024
pipe size (512 bytes, -p) 8
POSIX message queues (bytes, -q) 819200
real-time priority (-r) 0
stack size (kbytes, -s) 8192
cpu time (seconds, -t) unlimited
max user processes (-u) 94832
virtual memory (kbytes, -v) unlimited
file locks (-x) unlimited
I did just notice that it's showing 4gb less RAM than there should be (12 instead of 16, I'll be checking into that about the same time I hit submit) but as far as I can tell the system shouldn't even be worried about swap 99.99% of the time. Any ideas?
Edit: On reboot system show the full 16gb of RAM.
Last edited by inquisitorthreefive (2012-11-19 07:44:32)

inquisitorthreefive wrote:
I did just notice that it's showing 4gb less RAM than there should be (12 instead of 16, I'll be checking into that about the same time I hit submit) but as far as I can tell the system shouldn't even be worried about swap 99.99% of the time. Any ideas?
Edit: On reboot system show the full 16gb of RAM.
Could it be a hardware problem with RAM? Is there a 4 GB RAM chip in the machine that may have failed?

How to mount and use swap partition _only_ for hibernation ?

By "hibernation" I mean suspend to disk.
My disk is a bit poor in RAM and the dirsk is slow, so I want to prevent any swap use of my system.
AND I want to use hibernation, in order to boot fatser... and not to loose my workspace state.
With a swap partition, hibernation works nicely.
But any intensive web browser easily lealds to incresed memory usage, hence swap use.
I have lowered the swap priority in /etc/fstab, down to 1 (and even tried 0).
Still, I am not satisfied : If would like not to use swap at all but for hibernation.
My plan was hence :
1) disable the automatic swap mount at boot
2) add a systemd unit to only moutn before hibernation
The first part was easy : I added a "noauto" option in /etc/fstab. Great, no more mounted swap. But, as expected, no more successful hibernation.
Then I tried to create a custom systemd unit.
- first copy the unit file automatically generated by systemd in /run/systemd/generator/dev-....swap to /etc/systemd/system/
- and add "Before=sleep.target" line in the Unit section
Result file here :
> cat /etc/systemd/system/dev-disk-by\x2duuid-blablablabla.swap
[Unit]
Description=Swap required for hibernation
Before=sleep.target
[Swap]
What=/dev/disk/by-uuid/blablablablabla
Priority=5
>
Next step show be to enable it, but I first run a safety check :
> systemctl show dev-disk-by\x2duuid-blablablabla.swap
Id=dev-disk-by\x2duuid-blablablabla.swap
Names=dev-disk-by\x2duuid-blablablabla.swap
Requires=-.mount
Wants=system.slice
BindsTo=dev-disk-by\x2duuid-blablablabla.device
RequiredBy=swap.target
WantedBy=dev-disk-by\x2duuid-blablablabla.device
Conflicts=umount.target
Before=sleep.target umount.target swap.target
After=systemd-journald.socket dev-disk-by\x2duuid-blablablabla.device system.slice -.mount
RequiresMountsFor=/dev/disk/by-uuid/blablablabla
Description=Swap required for hibernation
LoadState=loaded
ActiveState=inactive
SubState=dead
FragmentPath=/etc/systemd/system/dev-disk-by\x2duuid-blablablabla.swap
UnitFileState=static
InactiveExitTimestampMonotonic=0
ActiveEnterTimestampMonotonic=0
ActiveExitTimestampMonotonic=0
InactiveEnterTimestampMonotonic=0
CanStart=yes
CanStop=yes
CanReload=no
CanIsolate=no
StopWhenUnneeded=no
RefuseManualStart=no
RefuseManualStop=no
AllowIsolate=no
DefaultDependencies=yes
>
The problem here is that, somehow, a RequiredBy=swap.target line is apaprently automatically added, and makes exactly NOT what I want.
I've been battling a bit with systemd, but so far I have no clue how to do this correctly.
Important : I wand a clean "official" way, not to do a "service" unit that calls an home-brewed script taht calls swapon. I can do that, but systemd should be able to do it better than me, and without any trick.

Hello and thank you all
@fdservices & @WonderWoofy :
In fact, I forgot to mentionned swappiness, just tought about mentionneing the swap priority which act at a very different level is only revelevant when you ahve several swap files.
I have already modified that setting since quite a long time... and it indeed makes the system behaviour so much better.
Honestly, I could do with it... that's what I am doing since a few months in fact, and it really largely improved the user experience.
Still, the setting only makes swapping unlikely, but does not forbid it.
@WonderWoofy & @lucke :
You may be right, I could satisfy with swapiness...
Still, my feeling is that swap at run-time is mostly a remeniscence of the time where having several megabytes of RAM was juste a fantasm. At that time, at least for some intensive operations, simulating more RAM thansk to the cheaper & slower disk capacity was really decisive for a good user experience... and to allow some operations like some huge compilations. My feeling is that today, with our Gigabytes of RAM on our desktop/laptop, the need is much lower. In fact a swap is probably counter-productive for SSD for example (fast, yes, but swap is by nature something you write quite often, reducing teh SSD lifetime significantly). And simulating extra RAM is also counter-productive with applications such as web browsers which handle large amount of memory to handle cache themselves; it causes memory pressure, hence trigger swapping and machien slowdown instead of trigerring their pseudo-smart memory cleaning mechanisms. Last but not least, it does not encourage any improvement on application memory needs and rationalisation.
In fact, unless when I compile huge programs, or on really-limited-memory (embedded) systems, or eventually if swap can be hosted on significantly faster-than-storage disks, I feel that swapping does not really improve the user experience and system performance... and in contrary reduces them by having to handle swapping on and off memory & cache pages.
SO that's why I want to try. It's OK if I fail, swappinees fixes most my user experience issue, but I try to solve my more "theorical" issue as well.
And honestly, it is also to lean & play a bit with systemd as well.
More to come, I have things partially working now. Partially for now, and maybe perfectly with your help ?

[SOLVED] Swap Partition keeps being automounted, but it shouldn't.

Hi,
I created a swap partition as part of setting up a Linux From Scratch installation, and it was intended for that system only. However it is automounted every time I boot my arch.
It is neither mentioned in my fstab or my mtab. I tried # swapoff , and that disables it for that instance, but after boot it's mounted again. Tried giving it an fstab entry with the noauto option.
Last edited by krork (2015-01-05 18:11:24)

Thanks Guys!
I have not yet found a way to stop systemd from automounting the swap partition. I'm a bit busy this January. However, as a workaround I created a swapoff.service file in /etc/systemd/system with this content:
[Unit]
Description=Turns off Swap Partition
DefaultDependencies=no
After=getty.target
[Service]
ExecStart=/usr/bin/swapoff /dev/sdaX
[Install]
WantedBy=multi-user.target
Replace the /dev/sdaX with your according partition, of course.
Edit: Also you'll want to enable it: # systemctl enable swapoff.service
So in case anyone else is having this problem and just wants a dirty fix, this worked for me, for now. If I find out how to keep systemd from automounting in the first place I'll let you know.
Last edited by krork (2015-01-05 16:50:33)

No swap partition on 10g DB

I had a doubt that is a partition with the name 'swap' needs to exist on a 10g DB server for swap space to be used. My current 10g DB server does not have a partition with the name 'swap' explicitly. It has the following kind of structure.
[root@JispNewDB scripts]# df -kh
Filesystem Size Used Avail Use% Mounted on
/dev/cciss/c0d0p8 48G 12G 34G 26% /
/dev/cciss/c0d0p3 48G 22G 24G 49% /backup
/dev/cciss/c0d0p1 99M 12M 83M 12% /boot
/dev/cciss/c0d0p2 191G 33G 149G 18% /crestel
none 4.0G 0 4.0G 0% /dev/shm
/dev/cciss/c0d0p9 24G 81M 23G 1% /home
/dev/cciss/c0d0p5 48G 6.6G 39G 15% /indexes
/dev/cciss/c0d0p7 48G 85M 46G 1% /tmp
/dev/cciss/c0d0p6 48G 32G 14G 70% /u01
Does, it mean that swap space will not be used in the system? There is 16 GB of swap space otherwise on the system.
I hope, my question is clear.
Please, help in solving the doubt.
regards

Is command free showing swap?
If You have linux, then You can execute
swapon -s - this will show all swap files or partition defined on system
look into /etc/fstabs as well to see is there defined swap partition.
fdisk -ls will show as well.
If above are showing swap - then You are using swap.

[SOLVED] Please verify: swap partition needed for dm-crypt/LUKS??

Hi,
Can anyone confirm whether or not the following is accurate (from http://wiki.archlinux.org/index.php/Sys … artitions)
There are 3 required partitions for any encrypted system:
- The root file system: /
- The initial boot partition: /boot
- The swap partition: swap
Do I need a swap partition if I encrypt? I can't seem to find reference to swap being mandatory anywhere else.
Last edited by jwhendy (2010-10-12 13:27:12)

You don't need swap. The minimal required setup is having two partitions: boot and root, where /boot must not be encrypted and / can be encrypted.
Last edited by stfn (2010-10-12 12:45:22)

Arch is trying to activate my swap partition twice

It's not really a serious problem at the moment, but it is annoying seeing an error on every boot. Arch is trying to activate my swap partition twice, once by UUID and once as /dev/sda7. I have explicitly specified in my fstab to have the swap partition activated by UUID, so I'm not sure what's causing it to also be mounted as /dev/sda7, but whichever of the two occurs second always fails. My fstab only has one entry per partition, and all are by UUID. Does anyone have any idea what else might be trying to activate the swap partition?
My fstab (mostly automatically generated, though I do think I remember removing one or two problem entries in the past):
# /dev/sda6
UUID=67d13305-c5d1-494c-a810-cf6eb3ac4924 / ext4 rw,relatime,data=ordered 0 1
# /dev/sda2 LABEL=ESP
UUID=B212-E288 /boot/efi vfat rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=iso8859-1,shortname=mixed,errors=remount-ro 0 2
# /dev/sda7
UUID=4d2464ab-4499-40be-948e-2261efa79396 none swap defaults 0 0
tmpfs /tmp tmpfs nodev,nosuid,size=4G 0 0

karol wrote:Funny, it systemd recently had a bug which didn't enable swap at all https://mailman.archlinux.org/pipermail … 34145.html ;P
Wasn't there a patch for it in 207 and it was fixed with 208???
I believe i have a similar situation.
Seems to me that systemd mounts it by pattuuid and then tries to mount it by uuid (but i have no knowledge in that stuff so i probably talk bullshit)
systemctl
dev-disk-by\x2dlabel-Swap.swap loaded active active /dev/disk/by-label/Swap
dev-disk-by\x2dpartlabel-Swap.swap loaded active active /dev/disk/by-partlabel/Swap
dev-disk-by\x2dpartuuid-6a4d9284\x2de456\x2d4150\x2d9fa6\x2d055acb042649.swap loaded active active /dev/disk/by-partuuid/6a4d9284-e456-4150-9fa6-055acb042649
dev-disk-by\x2duuid-8fa200a9\x2d4cd4\x2d4967\x2d853c\x2d887ee6cae652.swap loaded failed failed /dev/disk/by-uuid/8fa200a9-4cd4-4967-853c-887ee6cae652
dev-sda2.swap loaded active active /dev/sda2
journal
Oct 13 10:28:37 mainland systemd[1]: Activating swap /dev/sda2...
Oct 13 10:28:37 mainland systemd[1]: Activating swap /dev/disk/by-uuid/8fa200a9-4cd4-4967-853c-887ee6cae652...
Oct 13 10:28:37 mainland swapon[307]: swapon: /dev/disk/by-uuid/8fa200a9-4cd4-4967-853c-887ee6cae652: swapon failed: Device or resource busy
Oct 13 10:28:37 mainland systemd[1]: Activated swap /dev/sda2.
Oct 13 10:28:37 mainland systemd[1]: dev-disk-by\x2duuid-8fa200a9\x2d4cd4\x2d4967\x2d853c\x2d887ee6cae652.swap swap process exited, code=exited status=255
Oct 13 10:28:37 mainland systemd[1]: Failed to activate swap /dev/disk/by-uuid/8fa200a9-4cd4-4967-853c-887ee6cae652.
Oct 13 10:28:37 mainland systemd[1]: Dependency failed for Swap.
Oct 13 10:28:37 mainland systemd[1]: Unit dev-disk-by\x2duuid-8fa200a9\x2d4cd4\x2d4967\x2d853c\x2d887ee6cae652.swap entered failed state.
fstab
# /etc/fstab: static file system information
# <file system> <dir> <type> <options> <dump> <pass>
tmpfs /tmp tmpfs nodev,nosuid 0 0
UUID=33D6-79D6 /boot/efi vfat defaults 0 1
UUID=8fa200a9-4cd4-4967-853c-887ee6cae652 none swap defaults 0 0
UUID=447f9004-da86-4f8b-b92a-2cdd4fc5b61b / ext4 defaults,noatime,discard 0 1

Creating Swap Partitions for Oracle 8

I am in the delimma of having already installed RedHat 5.2, with
the intention of installing Oracle 8, however, I have created a
swap partition of 32MB on a machine that has 48MB of RAM.
Rather than repartition my hard drive to create a swap partition
that is three times the size of the RAM installed, (as
recommended by the installation guide), would I be able to
create a swap file to overcome this limitation?
null

Bill Tourloupis (guest) wrote:
: I am in the delimma of having already installed RedHat 5.2,
with
: the intention of installing Oracle 8, however, I have created a
: swap partition of 32MB on a machine that has 48MB of RAM.
: Rather than repartition my hard drive to create a swap
partition
: that is three times the size of the RAM installed, (as
: recommended by the installation guide), would I be able to
: create a swap file to overcome this limitation?
You can create swap on a file using the following script. Replace
"u01" with your mount point. Change "count" to the number of 1K
blocks you want.
dd if=/dev/zero of=/u01/swapfile bs=1024 count=524288
mkswap -c -v1 /u01/swapfile
chmod 0600 /u01/swapfile
swapon /u01/swapfile
swapon -s
null

SWAP partition (Hibernate Failure)[SOLVED]

Hi Folks,
My suspend to disk (Hibernate) is not working. When I boot up my pc, it starts processing. In meanwhile, I can see the output on my screen. At some point, it starts checking mounted filesystems
"Checking Filesystems"
/dev/sda1
dev/sda3
/dev/sda4
I do not know why this skips /dev/sda2 which is my swap partition actually.
After three or four lines, I can see from the output saying "Activating Swap". Does it mean that my swap partition is configured correctly or not?
There is one more thing. I use gnome. In gnome, when I open System Monitor (from the panel) > and under File Systems, I also can not see my swap partition which is /dev/sda2 while I can see all other partitions which are /, /boot, /home.
Does it mean that the partition can not be mounted correctly.
I really appreciate any kind of help.
Thanks in advance.
Last edited by Archie_Enthusiasm (2010-11-28 15:49:30)

Here is the output:
cat /proc/swaps
Filename Type Size Used Priority
/dev/sda2 partition 6297476 0 -1
free -m
total used free shared buffers cached
Mem: 3013 1182 1830 0 58 532
-/+ buffers/cache: 591 2421
Swap: 6149 0 6149
It says that the swap is being used. But when I hibernate my pc, it correctly shuts itself down. But when I wake up the pc again, it says at the boot time that my / , /root, /boot mount file systems were not unmounted correctly and starts checking the filesystems (Some kind of journaling process starts). When it boots up, the stored session is no more there or existent anymore. It is like as I forced the system to shutdown and start it again (Just like as there were a system freeze and then I did shutdown the system by pressing the power button for more than 10 secs and then restarted). So I suspect that the system did not shutdown correctly even if there were no strange occurrence happening when I did the hibernate.
I hope someone can help me. If you need some more information about the configuration on my pc, please tell me command and I will post the output.
Thanks in advance
Last edited by Archie_Enthusiasm (2010-10-28 21:41:58)

Is a swap partition necessary?

I was wondering if the Arch install procedure allows for
installing without creating a swap partition. In my current hard
disk setup it would be simpler if I did not have to change
the partitioning.
Creating a swap file instead on an existing partition works
just the same, as far as I can tell.
P.S. Does the base CD come with a PPPOE dialer?
Thanks to all for any help.

segmental wrote:I was wondering if the Arch install procedure allows for
installing without creating a swap partition.
of course.
If you already have prepared partitions (or aren't afraid doing this by hand) I would recommend you quickinstall script that comes with iso image.
Check wiki.archlinux.org for detailed instruction on how to use it.
cheers

Performance: swap space vs. swap file?

After install of Red Hat Linux ES 3.0 UL3, I found that the swap space is set to only 2 GB. The SA claims that to increase the swap space, he must reinstall Linux because the partition is full.
Another option would be to add a swap file to a free partition.
Are there any issues with having swap files vs. swap space?

You should be able to use the free partition as swap space as well, without reinstall of Linux.
More information is available from the Redhat website :
http://www.redhat.com/docs/manuals/linux/RHL-8.0-Manual/custom-guide/s1-swap-adding.html

Please make a swap partition optional at installation

Because i have enough RAM i would like to skip the creation of a swap partition at the setup.
Could you please make this available at the next (0.7) release?
lubot

As far as I remember Con Kolivas was testing some VM managers lately in his ck kernels.
These include hard swapiness (in <2.6.8.1-ck4) and mapped watermark (in 2.6.8.1-ck4).
http://members.optusnet.com.au/ckolivas/kernel/
Sun Aug 22 20:19:47 EST 2004
Mapped watermark??
This readjusts the way memory is evicted by lightly removing cached ram once the ram is more than 2/3 full, if less than the "mapped watermark" percent of ram is mapped ram (ie applications). The normal system is to aggresively start scanning ram once it is completely full. The benefits of this are:
1. Allocating memory while ram is being lightly scanned is faster and cheaper than when it is being heavily scanned.
2. There is usually some free ram which tends to speed up application startup times.
3. Swapping is an unusual event instead of a common one if you have enough ram for your workload.
4. It is rare for your applications to be swapped out by file cache pressure.
Disadvantage: Less file cache.
The mapped watermark is configurable so a server for example might be happy to have a lower mapped percentage. The default is 66 and a server might like 33
echo 33 > /proc/sys/vm/mapped
This removes the swappiness knob entirely and deprecates all my previous vm hacks (autoregulated swappiness, hard swappiness, kiflush).
You must be aware that the normal system usually shrinks the memory allocated for buffers & file cache until it becomes close to ZERO, making almost all the memory available for user applications.

Resume from suspend requires bogus swap partition in suspend.conf

OK. This is all a bit weird but hopefully I can explain everything.
Suspend and resume actually work BUT ... I have to make /etc/suspend.conf resume device = 'something other than my swap partition'
This causes the boot process to halt with an error message saying ... "cannot stat resume device .." because the uresume hook has been added and the resume device is bogus (obviously!)
I can just press enter and the boot process continues but it's a bit annoying to have to do this every time I want to boot and of course this shouldn't be necessary.
If I make /etc/suspend.conf resume device = /dev/sda5 which is a swap partition that I only just created as per all the instructions on wiki pages (despite not understanding why suspend needs it) then suspend still works but resume doesn't. The disk spins up, the lights go on as does wireless etc but the screen has black and white stripes and the system is absolutely unresponsive. I have to hard reboot.
I have gone through all the wiki pages on suspend, suspend to ram etc etc.
I'm using uswsusp.
Prior to the latest few weeks of package updates, suspend and resume worked without me ever doing a single thing. No editing of files, not uresume hooks, resume=/dev/sda5 in grub nothing. I never even knew /etc/pm/config.d/config existed because I literally did NOTHING. It just worked.
Any help would be VERY much appreciated.
Last edited by tim (2009-07-23 13:46:54)

Where is you backup in sucha a case?

Swap Partition Vs Swap file

Similar Messages

Maybe you are looking for