[gentoo-user] syslog-ng: reporting no space though there's plenty

[gentoo-user] syslog-ng: reporting no space though there's plenty

[hw]

Hi,

syslog-ng keeps reporting:


Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has 
elapsed, attempting to write again; fd='15'
Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation 
because of an I/O error; fd='15', time_reopen='60'


while there's plenty of disk space available.  Restarting it didn't 
help.  This is from an LXC container --- could there be some disk limit 
in effect by default which I don't know about?

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[Fernando Rodriguez]

On Monday, March 23, 2015 5:37:28 PM hw wrote:
> 
> Hi,
> 
> syslog-ng keeps reporting:
> 
> 
> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has 
> elapsed, attempting to write again; fd='15'
> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation 
> because of an I/O error; fd='15', time_reopen='60'
> 
> 
> while there's plenty of disk space available.  Restarting it didn't 
> help.  This is from an LXC container --- could there be some disk limit 
> in effect by default which I don't know about?
> 

Post the output of:

ls -l /proc/27532/fd/15

where 27532 is the pid next syslog-ng on the log and 15 is the fd value.

-- 
Fernando Rodriguez

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[Alan McKinnon]

On 23/03/2015 18:37, hw wrote:
> 
> Hi,
> 
> syslog-ng keeps reporting:
> 
> 
> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has
> elapsed, attempting to write again; fd='15'
> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation
> because of an I/O error; fd='15', time_reopen='60'
> 
> 
> while there's plenty of disk space available.  Restarting it didn't
> help.  This is from an LXC container --- could there be some disk limit
> in effect by default which I don't know about?
> 


You are on the wrong track. That error message does not say there is no
space available.

It says the disk is not available, something very different. Check dmesg.

-- 
Alan McKinnon
alan.mckinnon@gmail.com

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[hw]

Am 23.03.2015 um 19:27 schrieb Fernando Rodriguez:
> ls -l /proc/27532/fd/15

sunflo-mx ~ # ls -l /proc/27532/fd/15
l-wx------ 1 root root 64 24. Mär 10:06 /proc/27532/fd/15 -> /dev/tty12
sunflo-mx ~ #


Cool, thank you, I think I understand :)  The host itself is using tty12 
with its own syslog-ng to put log output to ...  So I guess I should 
turn this off for the container.

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[hw]

Am 24.03.2015 um 08:01 schrieb Alan McKinnon:
> On 23/03/2015 18:37, hw wrote:
>>
>> Hi,
>>
>> syslog-ng keeps reporting:
>>
>>
>> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has
>> elapsed, attempting to write again; fd='15'
>> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation
>> because of an I/O error; fd='15', time_reopen='60'
>>
>>
>> while there's plenty of disk space available.  Restarting it didn't
>> help.  This is from an LXC container --- could there be some disk limit
>> in effect by default which I don't know about?
>>
>
>
> You are on the wrong track. That error message does not say there is no
> space available.
>
> It says the disk is not available, something very different. Check dmesg.
>

Mar 24 10:05:01 sunflo-mx syslog-ng[27532]: internal() messages are 
looping back, preventing loop by suppressing all internal messages until 
the current message is processed; trigger-msg='', 
first-suppressed-msg='I/O error occurred while writing; fd=\'15\', 
error=\'No space left on device (28)\''


Somewhat misleading maybe ...

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[Fernando Rodriguez]

On Tuesday, March 24, 2015 9:01:37 AM Alan McKinnon wrote:
> On 23/03/2015 18:37, hw wrote:
> > 
> > Hi,
> > 
> > syslog-ng keeps reporting:
> > 
> > 
> > Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has
> > elapsed, attempting to write again; fd='15'
> > Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation
> > because of an I/O error; fd='15', time_reopen='60'
> > 
> > 
> > while there's plenty of disk space available.  Restarting it didn't
> > help.  This is from an LXC container --- could there be some disk limit
> > in effect by default which I don't know about?
> > 
> 
> 
> You are on the wrong track. That error message does not say there is no
> space available.
> 
> It says the disk is not available, something very different. Check dmesg.
> 
> 

It just says that a write failed. It could be anything.
-- 
Fernando Rodriguez

Re: [gentoo-user] syslog-ng: reporting no space though there's plenty

[Fernando Rodriguez]

On Tuesday, March 24, 2015 10:09:50 AM hw wrote:
> 
> Am 24.03.2015 um 08:01 schrieb Alan McKinnon:
> > On 23/03/2015 18:37, hw wrote:
> >>
> >> Hi,
> >>
> >> syslog-ng keeps reporting:
> >>
> >>
> >> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has
> >> elapsed, attempting to write again; fd='15'
> >> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation
> >> because of an I/O error; fd='15', time_reopen='60'
> >>
> >>
> >> while there's plenty of disk space available.  Restarting it didn't
> >> help.  This is from an LXC container --- could there be some disk limit
> >> in effect by default which I don't know about?
> >>
> >
> >
> > You are on the wrong track. That error message does not say there is no
> > space available.
> >
> > It says the disk is not available, something very different. Check dmesg.
> >
> 
> Mar 24 10:05:01 sunflo-mx syslog-ng[27532]: internal() messages are 
> looping back, preventing loop by suppressing all internal messages until 
> the current message is processed; trigger-msg='', 
> first-suppressed-msg='I/O error occurred while writing; fd=\'15\', 
> error=\'No space left on device (28)\''
> 
> 
> Somewhat misleading maybe ...

I hate it when I run out of space on tty12 :)


-- 
Fernando Rodriguez

Re: [gentoo-user] syslog-ng: (solved) reporting no space though there's plenty

[hw]

Am 24.03.2015 um 10:13 schrieb Fernando Rodriguez:
> On Tuesday, March 24, 2015 9:01:37 AM Alan McKinnon wrote:
>> On 23/03/2015 18:37, hw wrote:
>>>
>>> Hi,
>>>
>>> syslog-ng keeps reporting:
>>>
>>>
>>> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Error suspend timeout has
>>> elapsed, attempting to write again; fd='15'
>>> Mar 23 17:34:41 sunflo-mx syslog-ng[27532]: Suspending write operation
>>> because of an I/O error; fd='15', time_reopen='60'
>>>
>>>
>>> while there's plenty of disk space available.  Restarting it didn't
>>> help.  This is from an LXC container --- could there be some disk limit
>>> in effect by default which I don't know about?
>>>
>>
>>
>> You are on the wrong track. That error message does not say there is no
>> space available.
>>
>> It says the disk is not available, something very different. Check dmesg.
>>
>>
>
> It just says that a write failed. It could be anything.
>

I think disabling the logging to the console for the container solves 
this; I commented out the entries in /etc/syslog-ng/syslog-ng.conf:


# destination console_all { file("/dev/tty12");
# log { source(src); destination(console_all);


... and restarted syslog-ng, and the error hasn't come back yet.