How to safely remove swap partition in Red Hat Enterprise Linux?
SOLUTION 已验证 - 已更新 2019年八月21日17:51 -
环境
- Red Hat Enterprise Linux 5
- Red Hat Enterprise Linux 6
- Red Hat Enterprise Linux 7
- Swap
问题
-
Is there a way to stop active swap disks, so that they can move their load to other swap disks?
-
Server being migrated from old SAN arrays to a new array. Secondary swap is currently in use as swap disks on the old array and the new array.
-
From below list, I want to stop sda2 and sda3, but they have a multi-threaded load with primary swap. I'm assuming that only a reboot will let me drop them (already commented out in fstab).
# swapon -s
Filename Type Size Used Priority
/dev/mapper/VolGroup00-LogVol03 partition 2097144 368892 2
/dev/sda1 partition 28846072 368520 2
/dev/sda2 partition 28846072 368364 2
/dev/sdj1 partition 29360120 0 -1
/dev/sdj2 partition 29360120 0 -2
决议
Swap space in Linux is used when the amount of physical memory (RAM) is full. If the system needs more memory resources and the RAM is full, inactive pages in memory are moved to the swap space. For more details refer Swap Space .
As per the swapon
output mentioned in issue section, other partitions have same partition size and priorities set. The swap partitions can simply get off by using following command and their load will get split.
# swapoff /dev/sda1
Check the load by using swapon -s
. Some of load will get split in memory and other in swap space. When it will restore and use the other swap partition, remove this partition /dev/sda1
using fdisk.
After doing this wait for some time to get it restore and use the other partition and repeat the same steps for device /dev/sda2
and remove the entries from /etc/fstab
.
After memory is swapped out, it will not return to the physical memory unless it is being used again. The pages can exist in both the places swap and main memory.
After swapoff
is issued, it will take some time to move the memory from swap partition to other partition and physical memory. Ensure enough memory is available otherwise the OOM (Out of memory) killer will be invoked.