Menu
▾
▴
Re: [Astlinux-devel] Adding a "spinner" indicator when e2fsck is working
|
From: Lonnie A. <li...@lo...> - 2017-08-29 18:30:52
|
On Aug 29, 2017, at 1:18 PM, Michael Keuter <li...@mk...> wrote:
>
>> Am 29.08.2017 um 20:12 schrieb Lonnie Abelbeck <li...@lo...>:
>>
>> Hi Devs,
>>
>> David made a good point, so I looked into piping the output of e2fsck to a "spinner" shell function, triggered by a newline and update no more often then ever second ...
>> --
>> spinner()
>> {
>> local i sp secs nsecs line IFS
>>
>> i=0
>> secs=0
>>
>> unset IFS
>> while read line; do
>> nsecs=$(cut -d. -f1 /proc/uptime)
>> if [ $nsecs -ne $secs ]; then
>> secs=$nsecs
>> i=$(((i+1)%4))
>> case $i in
>> 1) sp='-' ;;
>> 2) sp='\' ;;
>> 3) sp='|' ;;
>> *) sp='/' ;;
>> esac
>> echo -ne " Working[${sp}] \r"
>> fi
>> done
>>
>> echo -ne " \r"
>> }
>> --
>> It took some work getting this to work without printf, no date, no bash ${str:1:1}, etc., but it works with the current Busybox configs.
>>
>> But ... here are the two places "spinner" would apply to ...
>>
>> 1) in the initrd, linuxrc ...
>> --
>> e2fsck -y $ASTURW >/dev/null
>> status=$?
>>
>> case $status in
>> --
>>
>> 2) in astlinux, /etc/rc ...
>> --
>> echo "Checking $1"
>> e2fsck -y $1 >/dev/null
>>
>> status="$?"
>>
>> case $status in
>> --
>>
>> The plan would be to use:
>> --
>> e2fsck -y $1 | spinner
>> --
>> But the return value of $? is no longer from e2fsck, but rather spinner
>>
>> You might try ...
>> --
>> ( e2fsck -y $1 ; status="$?" ) | spinner
>> --
>> Almost, but now "status" is defined in a subshell which is not propagated up and lost.
>>
>> A simple but ugly solution is to use a temp file ...
>> --
>> ( e2fsck -y $1 ; echo "$?" > /tmp/e2fsck_status ) | spinner
>> status="$(cat /tmp/e2fsck_status)"
>> rm /tmp/e2fsck_status
>> --
>> but not to mention the initrd does not currently mount a /tmp filesystem. We would have to have to create and delete a tmpfs filesystem just to get the return value from e2fsck. :-(
>>
>> I'm leaning toward leaving things as is, comments ?
>>
>> Lonnie
>>
>>
>> On Aug 29, 2017, at 8:20 AM, Lonnie Abelbeck <li...@lo...> wrote:
>>
>>> Hi David,
>>>
>>> Agreed, no e2fsck output when it is forced to work for a long period is not ideal. Thankfully this is a very rare occurrence.
>>>
>>> If we can come up with a clever spinning wheel or such counting the lines e2fsck generates might be a good addition.
>>>
>>> Recall this all occurs in the "initrd", so the available packages and BusyBox config is quite limited.
>>>
>>> Any further discussion on this idea should move to the astlinux-devel list.
>>>
>>> Lonnie
>>>
>>>
>>> On Aug 29, 2017, at 7:48 AM, David Kerr <Da...@Ke...> wrote:
>>>
>>>> Is that a good idea (redirecting e2fsck to null)? 15 minutes is a long time to wait with no indication of anything happening.
>>>>
>>>> David
>>>>
>>>> On Tue, Aug 29, 2017 at 8:33 AM, Lonnie Abelbeck <li...@lo...> wrote:
>>>> Hi Tim,
>>>>
>>>> Yes, it was running "e2fsck" with stdout redirected to /dev/null so you did not see it working ... eventually e2fsck will return with a result code if left long enough.
>>>>
>>>> Good to hear your filesystem is now clean, you might reboot once again to make sure the filesystem is good.
>>>>
>>>> BTW, if the automatic e2fsck repair did not work, here is the manual fallback procedure:
>>>> --
>>>> ## reboot, and quickly when the RUNNIX boot menu appears type "shell"
>>>> boot: shell
>>>>
>>>> ## Wait for a "runnix# " CLI prompt.
>>>> ## determine the first Linux ext2 partition, usually always /dev/sda2
>>>> runnix# findfs LABEL=ASTURW
>>>>
>>>> ## using the findfs result, run e2fsck manually, you may want to add -y or -p options
>>>> runnix# e2fsck /dev/sda2
>>>>
>>>> ## output a list of options for e2fsck
>>>> runnix# e2fsck
>>>> Usage: e2fsck [-panyrcdfktvDFV] [-b superblock] [-B blocksize]
>>>> [-l|-L bad_blocks_file] [-C fd] [-j external_journal]
>>>> [-E extended-options] [-z undo_file] device
>>>>
>>>> Emergency help:
>>>> -p Automatic repair (no questions)
>>>> -n Make no changes to the filesystem
>>>> -y Assume "yes" to all questions
>>>> -c Check for bad blocks and add them to the badblock list
>>>> -f Force checking even if filesystem is marked clean
>>>> -v Be verbose
>>>> -b superblock Use alternative superblock
>>>> -B blocksize Force blocksize when looking for superblock
>>>> -j external_journal Set location of the external journal
>>>> -l bad_blocks_file Add to badblocks list
>>>> -L bad_blocks_file Set badblocks list
>>>> -z undo_file Create an undo file
>>>>
>>>> ## Depending on how you initially configured AstLinux, check if you also have a ASTKD partition
>>>> ## (typically /dev/sda3 if it exists)
>>>> runnix# findfs LABEL=ASTKD
>>>> ## If the ASTKD label exists repeat the e2fsck steps above using the ASTKD label partition.
>>>>
>>>> ## reboot by issuing "exit"
>>>> runnix# exit
>>>> --
>>>>
>>>> Lonnie
>>>>
>>>>
>>>> On Aug 29, 2017, at 7:02 AM, Tim Turpin <tt...@z-...> wrote:
>>>>
>>>>> Never mind. After sitting at that point for about 15 minutes, it somehow recovered and finished booting up. All seems well now.
>>>>>
>>>>> From: Tim Turpin [mailto:tt...@z-...]
>>>>> Sent: Tuesday, August 29, 2017 7:34 AM
>>>>> To: ast...@li...
>>>>> Subject: [Astlinux-users] ASTLinux stopped booting
>>>>>
>>>>> We took a power hit on our test system yesterday, and now it will only load up to a certain point and stops with the following screen:
>
> Lonnie, don't give up so quickly :-).
> What you have already is quite a good starting point.
> What about adding "something needed" to our initrd busybox.config …
>
> Michael
The initrd busybox config is not the problem, this is a fundamental problem using shell, though easily solved with bash, but bash is not a practical option with our initrd.
Get exit status of process that's piped to another
https://unix.stackexchange.com/questions/14270/get-exit-status-of-process-thats-piped-to-another
I dont see an elegant solution.
My only other thought would be to create a background "spinner" process, then call "e2fsck -y $1 >/dev/null" and status="$?" then stop the background spinner process.
Lonnie
|
