Add support to convert the VFAT file system on Arista switches to EXT4 #201
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,170 @@ | ||
| #!/bin/sh | ||
|
|
||
| case $1 in | ||
| prereqs) | ||
| exit 0 | ||
| ;; | ||
| esac | ||
|
|
||
| set -e | ||
| # set -x | ||
| total_mem=$(free | awk '/^Mem:/{print $2}') | ||
| tmpfs_size=$(( $total_mem / 20 * 17 )) | ||
| free_mem_thres=$(( $total_mem / 20 * 18 )) | ||
| tmp_mnt='/mnt/ramdisk-convfs' | ||
| root_mnt='/mnt/root-convfs' | ||
| root_dev='' | ||
| flash_dev='' | ||
| block_flash='' | ||
| aboot_flag='' | ||
| backup_file='' | ||
|
|
||
| # Get the fullpath of flash device, e.g., /dev/sda | ||
| get_flash_dev() { | ||
| for dev in $(ls /sys/block); do | ||
| local is_mmc=$(echo "$dev" | grep 'mmcblk.*boot.*' | cat) | ||
| if [ -n "$is_mmc" ]; then | ||
| continue | ||
| fi | ||
| local devid=$(realpath "/sys/block/$dev/device") | ||
| local is_device=$(echo "$devid" | grep '^/sys/devices/' | cat) | ||
| local is_flash=$(echo "$devid" | grep "$block_flash" | cat) | ||
| if [ -n "$is_device" -a -n "$is_flash" ]; then | ||
| flash_dev="/dev/$dev" | ||
| return 0 | ||
| fi | ||
| done | ||
| return 1 | ||
| } | ||
|
|
||
| # Wait for root_dev to be ready | ||
| wait_for_root_dev() { | ||
| local try_rounds=30 | ||
| while [ $try_rounds -gt 0 ]; do | ||
| if [ -e "$root_dev" ]; then | ||
| return 0 | ||
| fi | ||
| sleep 1 | ||
| try_rounds=$(( $try_rounds - 1 )) | ||
| done | ||
| return 1 | ||
| } | ||
|
|
||
| # Alway run cleanup before exit | ||
| cleanup() { | ||
| if grep -q "$root_mnt" /proc/mounts; then | ||
| umount "$root_mnt" | ||
| fi | ||
| if grep -q "$tmp_mnt" /proc/mounts; then | ||
| umount "$tmp_mnt" | ||
| fi | ||
| [ -e "$root_mnt" ] && rmdir "$root_mnt" | ||
| [ -e "$tmp_mnt" ] && rmdir "$tmp_mnt" | ||
| } | ||
| trap cleanup EXIT | ||
|
|
||
| notification() { | ||
| cat << EOF | ||
| A failure happend in modifying the root file system which stopped the upgrade. Manual interventions are needed to fix the issue. Note that: | ||
| 1) files in the old root file system may have been lost and the old partition table may have been corrupted; | ||
| 2) The files in the old root file system were copied to $tmp_mnt; | ||
| 3) The old partition table was dumped to the file $tmp_mnt/$backup_file by sfdisk; | ||
| 4) Quitting the current shell will lose all files mentioned above permanently. | ||
| EOF | ||
| } | ||
|
|
||
| run_cmd() { | ||
| if ! eval "$1"; then | ||
| echo "$2" | ||
| notification | ||
| sh | ||
| exit 1 | ||
| fi | ||
| } | ||
|
|
||
| # Extract kernel parameters | ||
| set -- $(cat /proc/cmdline) | ||
|
There was a problem hiding this comment. what does 'set --' mean? #Closed There was a problem hiding this comment. set is parsing = from a string. |
||
| for x in "$@"; do | ||
| case "$x" in | ||
| block_flash=*) | ||
| block_flash="${x#block_flash=}" | ||
| ;; | ||
| Aboot=*) | ||
| aboot_flag="${x#Aboot=}" | ||
| esac | ||
| done | ||
| root_dev="$ROOT" | ||
|
|
||
| #Check aboot and root_dev is vfat | ||
| [ -z "$aboot_flag" ] && exit 0 | ||
| if [ -z "$root_dev" ]; then | ||
| echo "Error: root device name is not provided" | ||
| exit 1 | ||
| fi | ||
| if ! wait_for_root_dev; then | ||
| echo "Error: timeout in waiting for $root_dev" | ||
| exit 1 | ||
| fi | ||
| blkid | grep "$root_dev.*vfat" -q || exit 0 | ||
|
There was a problem hiding this comment. Could you also filter the specific partition label, or other things indicating Arista prepared partition? I think the 'grep' is too loose here. #Closed There was a problem hiding this comment. This is actually done in our script. |
||
|
|
||
|
|
||
| # Get flash dev name | ||
| if [ -z "$block_flash" ]; then | ||
| echo "Error: flash device info is not provided" | ||
| exit 1 | ||
| fi | ||
| if ! get_flash_dev; then | ||
| echo "Error: flash device is not found" | ||
| exit 1 | ||
| fi | ||
|
|
||
| # Check memory size for tmpfs | ||
| free_mem=$(free | awk '/^Mem:/{print $4}') | ||
| if [ "$free_mem" -lt "$free_mem_thres" ]; then | ||
| echo "Error: memory is not enough" | ||
| exit 1 | ||
| fi | ||
|
|
||
| # Backup partition table | ||
| mkdir -p "$root_mnt" | ||
| mount "$root_dev" "$root_mnt" | ||
| backup_file=backup.$(date +%Y-%m-%d.%H-%M-%S) | ||
| sfdisk -d "$flash_dev" > "$root_mnt/$backup_file" | ||
|
|
||
| # Check total size of files in root | ||
| total_file_size=$(du -s "$root_mnt" | awk '{print $1}') | ||
| if [ "$total_file_size" -gt "$tmpfs_size" ]; then | ||
| echo "Error: total file size is too large" | ||
| exit 1 | ||
| fi | ||
|
|
||
| # Create tmpfs, and copy files to tmpfs | ||
| mkdir -p "$tmp_mnt" | ||
| mount -t tmpfs -o size="${tmpfs_size}k" tmpfs "$tmp_mnt" | ||
| cp -a "$root_mnt/." "$tmp_mnt/" | ||
| umount "$root_mnt" | ||
|
|
||
| #### Lines below will modify the root file system, so any failure will be trapped to shell for manual interventions. | ||
|
|
||
| # Create a new partition table (content in flash_dev will be deleted) | ||
| err_msg="Error: repartitioning $flash_dev failed" | ||
| cmd="echo ';' | sfdisk $flash_dev" | ||
| run_cmd "$cmd" "$err_msg" | ||
|
|
||
| sleep 5 | ||
| err_msg="Error: timeout in waiting for $root_dev after repartition" | ||
| cmd="wait_for_root_dev" | ||
| run_cmd "$cmd" "$err_msg" | ||
|
|
||
| err_msg="Error: formatting to ext4 failed" | ||
| cmd="mke2fs -t ext4 -m2 -F -O '^huge_file' $root_dev" | ||
| run_cmd "$cmd" "$err_msg" | ||
|
|
||
| err_msg="Error: mounting $root_dev to $root_mnt failed" | ||
| cmd="mount -t ext4 $root_dev $root_mnt" | ||
| run_cmd "$cmd" "$err_msg" | ||
|
|
||
| err_msg="Error: copying files form $tmp_mnt to $root_mnt failed" | ||
| cmd="cp -a $tmp_mnt/. $root_mnt/" | ||
| run_cmd "$cmd" "$err_msg" | ||
|
|
||
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,50 @@ | ||
| #!/bin/sh | ||
| #Part of the code is revised based on initramfs-tools/hooks/fsck and initramfs-tool is under GPL v2. | ||
|
|
||
| PREREQ="" | ||
|
|
||
| prereqs() | ||
| { | ||
| echo "$PREREQ" | ||
| } | ||
|
|
||
| case $1 in | ||
| prereqs) | ||
| prereqs | ||
| exit 0 | ||
| ;; | ||
| esac | ||
|
|
||
| . /usr/share/initramfs-tools/hook-functions | ||
|
|
||
| copy_exec /sbin/mke2fs | ||
| copy_exec /sbin/sfdisk | ||
| copy_exec /sbin/fdisk | ||
|
|
||
| fstypes="ext4" | ||
|
|
||
| for type in $fstypes; do | ||
| prog="/sbin/mkfs.${type}" | ||
| if [ -h "$prog" ]; then | ||
| link=$(readlink -f "$prog") | ||
| copy_exec "$link" | ||
| ln -s "$link" "${DESTDIR}/$prog" | ||
| elif [ -x "$prog" ] ; then | ||
| copy_exec "$prog" | ||
| else | ||
| echo "Warning: /sbin/mkfs.${type} doesn't exist, can't install to initramfs, ignoring." | ||
| fi | ||
| done | ||
|
|
||
| for type in $fstypes; do | ||
| prog="/sbin/fsck.${type}" | ||
| if [ -h "$prog" ]; then | ||
| link=$(readlink -f "$prog") | ||
| copy_exec "$link" | ||
| ln -s "$link" "${DESTDIR}/$prog" | ||
| elif [ -x "$prog" ] ; then | ||
| copy_exec "$prog" | ||
| else | ||
| echo "Warning: /sbin/fsck.${type} doesn't exist, can't install to initramfs, ignoring." | ||
| fi | ||
| done |
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are these parameters good for any Arsita switches? What is the rule behind the numbers?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The script need a memory-based tmpfs to backup and restore the files in the flash during the conversion. The size of created tmpfs is expected to be as large as possible to backup all files in the flash, but it is constrained by the free memory available.
In the script, tmpfs_size is the size of tmpfs to create, and we expect that the free memory space is at least free_mem_thres, which is tmpfs_size + some extra.
According to our testing, the script with the current settings can always create a tmpfs, and never fail at the check of free memory space (line 123). (The script runs as the early stage of booting, so an extremely large portion of the memory should be free.) A special case is that when tmpfs_size is smaller than the total size of files in the flash (line 136), as there is not enough space to backup all files, the booting will continue without any change to the file system and partition. (But this will never happen on the boxes with 2G flash and 4G memory.)