"Filesystem flags changed" alerts since 5.22

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

"Filesystem flags changed" alerts since 5.22

Paul Theodoropoulos
I'm getting lots of alerts for "Filesystem flags changed" since
immediately after updating to 5.22.

Debian 8.7

All of my machines have at least the following in their confs:

        check filesystem rootfs with path /
        if space greater 90 % then alert

No other types of filesystem tests are being performed. I noticed that
the docs say that the filesystem changed test is implicit. The problem
is that the filesystem hasn't changed. Here's the specific error message
generated by monit. These are all servers that have been up quite a while.

filesystem flags changed to rw,relatime,nobarrier,data=ordered

Am I doing something wrong?


--
Paul Theodoropoulos
www.anastrophe.com

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general
Reply | Threaded
Open this post in threaded view
|

Re: "Filesystem flags changed" alerts since 5.22

Paul Theodoropoulos
I should also add that it is only 'rootfs' that generates the error. If
a machine has additional mount points that are checked, they show no error.

On 4/19/17 15:47, Paul Theodoropoulos wrote:

> I'm getting lots of alerts for "Filesystem flags changed" since
> immediately after updating to 5.22.
>
> Debian 8.7
>
> All of my machines have at least the following in their confs:
>
>      check filesystem rootfs with path /
>      if space greater 90 % then alert
>
> No other types of filesystem tests are being performed. I noticed that
> the docs say that the filesystem changed test is implicit. The problem
> is that the filesystem hasn't changed. Here's the specific error message
> generated by monit. These are all servers that have been up quite a while.
>
> filesystem flags changed to rw,relatime,nobarrier,data=ordered
>
> Am I doing something wrong?
>
>

--
Paul Theodoropoulos
www.anastrophe.com

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general
Reply | Threaded
Open this post in threaded view
|

Re: "Filesystem flags changed" alerts since 5.22

Paul Theodoropoulos
Another addition, apologies for not having this all rolled into my
original message - I'm not getting alerts, only warnings within monit
summary, and m/monit.

On 4/19/17 16:45, Paul Theodoropoulos wrote:

> I should also add that it is only 'rootfs' that generates the error. If
> a machine has additional mount points that are checked, they show no error.
>
> On 4/19/17 15:47, Paul Theodoropoulos wrote:
>> I'm getting lots of alerts for "Filesystem flags changed" since
>> immediately after updating to 5.22.
>>
>> Debian 8.7
>>
>> All of my machines have at least the following in their confs:
>>
>>      check filesystem rootfs with path /
>>      if space greater 90 % then alert
>>
>> No other types of filesystem tests are being performed. I noticed that
>> the docs say that the filesystem changed test is implicit. The problem
>> is that the filesystem hasn't changed. Here's the specific error
>> message generated by monit. These are all servers that have been up
>> quite a while.
>>
>> filesystem flags changed to rw,relatime,nobarrier,data=ordered
>>
>> Am I doing something wrong?
>>
>>
>

--
Paul Theodoropoulos
www.anastrophe.com

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general
Reply | Threaded
Open this post in threaded view
|

Re: "Filesystem flags changed" alerts since 5.22

martinp@tildeslash.com
Hello Paul,

the problem may occur, if you monitor the root filesystem using the mountpoint ("/") and the system has an rootfs overlay like this:

        $ cat /proc/mounts
        rootfs / rootfs rw 0 0
        /dev/sda1 / ext3 rw,data=ordered 0 0

When monit scans the table, it finds the first rootfs mount with "rw" flags and then the second entry for /dev/sda1 with "rw,data=ordered" flags. This triggers the filesystem change event. The virtual rootfs filesystem is present on some platforms only (for example Ubuntu doesn't show it).

Workaround: use device instead of mountpoint for root filesystem monitoring (the device is more specific and doesn't match the whole overlay mount tree).

We have fixed the problem in the development version, if you want to test, you can compile it this way:

wget https://bitbucket.org/tildeslash/monit/get/master.tar.gz
tar -xzf master.tar.gz
cd tildeslash*
./bootstrap
./configure
make

Best regards,
Martin



> On 20 Apr 2017, at 02:04, Paul Theodoropoulos <[hidden email]> wrote:
>
> Another addition, apologies for not having this all rolled into my original message - I'm not getting alerts, only warnings within monit summary, and m/monit.
>
> On 4/19/17 16:45, Paul Theodoropoulos wrote:
>> I should also add that it is only 'rootfs' that generates the error. If a machine has additional mount points that are checked, they show no error.
>> On 4/19/17 15:47, Paul Theodoropoulos wrote:
>>> I'm getting lots of alerts for "Filesystem flags changed" since immediately after updating to 5.22.
>>>
>>> Debian 8.7
>>>
>>> All of my machines have at least the following in their confs:
>>>
>>>     check filesystem rootfs with path /
>>>     if space greater 90 % then alert
>>>
>>> No other types of filesystem tests are being performed. I noticed that the docs say that the filesystem changed test is implicit. The problem is that the filesystem hasn't changed. Here's the specific error message generated by monit. These are all servers that have been up quite a while.
>>>
>>> filesystem flags changed to rw,relatime,nobarrier,data=ordered
>>>
>>> Am I doing something wrong?
>>>
>>>
>
> --
> Paul Theodoropoulos
> www.anastrophe.com
>
> --
> To unsubscribe:
> https://lists.nongnu.org/mailman/listinfo/monit-general


--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general
Reply | Threaded
Open this post in threaded view
|

Re: "Filesystem flags changed" alerts since 5.22

Paul Theodoropoulos
Thanks very much, Martin. Setting it as the underlying device works
perfectly. I'm not sure if I'd even characterize this as a problem -
perhaps just a note in the docs is all that's needed...?


On 4/20/17 05:47, Martin Pala wrote:

> Hello Paul,
>
> the problem may occur, if you monitor the root filesystem using the mountpoint ("/") and the system has an rootfs overlay like this:
>
> $ cat /proc/mounts
> rootfs / rootfs rw 0 0
> /dev/sda1 / ext3 rw,data=ordered 0 0
>
> When monit scans the table, it finds the first rootfs mount with "rw" flags and then the second entry for /dev/sda1 with "rw,data=ordered" flags. This triggers the filesystem change event. The virtual rootfs filesystem is present on some platforms only (for example Ubuntu doesn't show it).
>
> Workaround: use device instead of mountpoint for root filesystem monitoring (the device is more specific and doesn't match the whole overlay mount tree).
>
> We have fixed the problem in the development version, if you want to test, you can compile it this way:
>
> wget https://bitbucket.org/tildeslash/monit/get/master.tar.gz
> tar -xzf master.tar.gz
> cd tildeslash*
> ./bootstrap
> ./configure
> make
>
> Best regards,
> Martin
>
>
>
>> On 20 Apr 2017, at 02:04, Paul Theodoropoulos <[hidden email]> wrote:
>>
>> Another addition, apologies for not having this all rolled into my original message - I'm not getting alerts, only warnings within monit summary, and m/monit.
>>
>> On 4/19/17 16:45, Paul Theodoropoulos wrote:
>>> I should also add that it is only 'rootfs' that generates the error. If a machine has additional mount points that are checked, they show no error.
>>> On 4/19/17 15:47, Paul Theodoropoulos wrote:
>>>> I'm getting lots of alerts for "Filesystem flags changed" since immediately after updating to 5.22.
>>>>
>>>> Debian 8.7
>>>>
>>>> All of my machines have at least the following in their confs:
>>>>
>>>>      check filesystem rootfs with path /
>>>>      if space greater 90 % then alert
>>>>
>>>> No other types of filesystem tests are being performed. I noticed that the docs say that the filesystem changed test is implicit. The problem is that the filesystem hasn't changed. Here's the specific error message generated by monit. These are all servers that have been up quite a while.
>>>>
>>>> filesystem flags changed to rw,relatime,nobarrier,data=ordered
>>>>
>>>> Am I doing something wrong?
>>>>
>>>>
>>
>> --
>> Paul Theodoropoulos
>> www.anastrophe.com
>>
>> --
>> To unsubscribe:
>> https://lists.nongnu.org/mailman/listinfo/monit-general
>

--
Paul Theodoropoulos
www.anastrophe.com

--
To unsubscribe:
https://lists.nongnu.org/mailman/listinfo/monit-general