GAE critical message: Exceeded soft memory limit
Jackie Fei
"Exceeded soft memory limit with 187.727 MB after servicing 1 requests
total"
My application has run on GAE for 50+ days, everything is smooth.
But I got the GAE critical message continuously from last night.
I cannot find any solution from global google search.
Could you give me any advices?
Thanks in advanced.
Jackie
Blake
haven't pushed any updates, so I don't believe it's an error on my
end.
Anyone else?
Jon McAlister
clones exceed the soft memory limit forever, but now we are letting
you know via logs. These logs can of course be ignored if you like,
since it doesn't effect the http response you served, but if you are
concerned about performance or latency, then I would pay attention to
these logs. The thing is, when a clone is killed after exceeding the
soft memory limit, it is dead and will receive no more requests. This
means that the next request for your app will have to be a loading
request, which depending on your app could mean it involves a much
higher latency and cpu consumption. Reducing the memory usage in your
clones could then mean you have fewer soft memory limit terminations,
and as such fewer loading requests.
> --
> You received this message because you are subscribed to the Google Groups "Google App Engine" group.
> To post to this group, send email to google-a...@googlegroups.com.
> To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.
>
>
Ikai L (Google)
application instance.
--
Ikai Lan
Developer Programs Engineer, Google App Engine
http://googleappengine.blogspot.com | http://twitter.com/app_engine
Jackie Fei
Thanks for your responses.
If this "critical" message is just a "notification", how could I
"hide" it?
Because my app is really need so huge memory to run smoothly and no
loading request performance issue.
These critical messages had flooded my debugging messages terribly.
Or put these critical messages elsewhere, or lower level than debug
message?
Thanks in advanced.
Jackie
On 3月26日, 上午1時59分, "Ikai L (Google)" <ika...@google.com> wrote:
> Minor clarification: Jon uses the word "clone" to refer to a running
> application instance.
>
>
>
>
>
> On Thu, Mar 25, 2010 at 10:51 AM, Jon McAlister <jon...@google.com> wrote:
> > This is a new logging message we just added. You've most likely had
> > clones exceed the soft memory limit forever, but now we are letting
> > you know via logs. These logs can of course be ignored if you like,
> > since it doesn't effect the http response you served, but if you are
> > concerned about performance or latency, then I would pay attention to
> > these logs. The thing is, when a clone is killed after exceeding the
> > soft memory limit, it is dead and will receive no more requests. This
> > means that the next request for your app will have to be a loading
> > request, which depending on your app could mean it involves a much
> > higher latency and cpu consumption. Reducing the memory usage in your
> > clones could then mean you have fewer soft memory limit terminations,
> > and as such fewer loading requests.
>
> > On Wed, Mar 24, 2010 at 9:19 PM, Blake <blake.commag...@gmail.com> wrote:
> >> I also started seeing this just last night for the first time. I
> >> haven't pushed any updates, so I don't believe it's an error on my
> >> end.
> >> Anyone else?
>
> >> On Mar 23, 7:23 pm, Jackie Fei <3s4m...@gmail.com> wrote:
> >>> Dear all,
>
> >>> "Exceeded soft memory limit with 187.727 MB after servicing 1 requests
> >>> total"
>
> >>> My application has run on GAE for 50+ days, everything is smooth.
> >>> But I got the GAE critical message continuously from last night.
>
> >>> I cannot find any solution from global google search.
> >>> Could you give me any advices?
>
> >>> Thanks in advanced.
>
> >>> Jackie
>
> >> --
> >> You received this message because you are subscribed to the Google Groups "Google App Engine" group.
> >> To post to this group, send email to google-a...@googlegroups.com.
> >> To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
> >> For more options, visit this group athttp://groups.google.com/group/google-appengine?hl=en.
>
> > --
> > You received this message because you are subscribed to the Google Groups "Google App Engine" group.
> > To post to this group, send email to google-a...@googlegroups.com.
> > To unsubscribe from this group, send email to google-appengi...@googlegroups.com.
> > For more options, visit this group athttp://groups.google.com/group/google-appengine?hl=en.
Eli Jones
notreadbyhumans
message. I have just started receiving it on what I thought was a
fairly minor operation (calling ~1,500 records from the datastore and
then looping through them to create a list based on various
properties). I now get the following message each time this process
runs:
Exceeded soft memory limit with 187.34 MB after servicing 1 requests
total
This seems very severe given the nature of the process that I am
running (according to the dashboard my entire datastore is only 31
MB!)
On Mar 26, 3:36 am, Eli Jones <eli.jo...@gmail.com> wrote:
> This is a nice addition.. I use a lot of per instance (or in clone cache or
> whatever one may call it) memory to cache datastore info (so.. the code
> checks its local cache.. then memcache and then datastore).. but I've been
> stuck sort of monitoring my tasks (I run looping tasks that access the same
> data) to see if I can tell when they hit the memory limit and do a cold
> start.. and then reduce the size of the in memory cache dict.
>
> Now, I'm guessing this will explicitly tell me when it's hitting that
> limit.. which will help a lot.
>
>
>
> On Thu, Mar 25, 2010 at 1:51 PM, Jon McAlister <jon...@google.com> wrote:
> > This is a new logging message we just added. You've most likely had
> > clones exceed the soft memory limit forever, but now we are letting
> > you know via logs. These logs can of course be ignored if you like,
> > since it doesn't effect the http response you served, but if you are
> > concerned about performance or latency, then I would pay attention to
> > these logs. The thing is, when a clone is killed after exceeding the
> > soft memory limit, it is dead and will receive no more requests. This
> > means that the next request for your app will have to be a loading
> > request, which depending on your app could mean it involves a much
> > higher latency and cpu consumption. Reducing the memory usage in your
> > clones could then mean you have fewer soft memory limit terminations,
> > and as such fewer loading requests.
>
> > On Wed, Mar 24, 2010 at 9:19 PM, Blake <blake.commag...@gmail.com> wrote:
> > > I also started seeing this just last night for the first time. I
> > > haven't pushed any updates, so I don't believe it's an error on my
> > > end.
> > > Anyone else?
>
> > > On Mar 23, 7:23 pm, Jackie Fei <3s4m...@gmail.com> wrote:
> > >> Dear all,
>
> > >> "Exceeded soft memory limit with 187.727 MB after servicing 1 requests
> > >> total"
>
> > >> My application has run on GAE for 50+ days, everything is smooth.
> > >> But I got the GAE critical message continuously from last night.
>
> > >> I cannot find any solution from global google search.
> > >> Could you give me any advices?
>
> > >> Thanks in advanced.
>
> > >> Jackie
>
> > > --
> > > You received this message because you are subscribed to the Google Groups
> > "Google App Engine" group.
> > > To post to this group, send email to google-a...@googlegroups.com.
> > > To unsubscribe from this group, send email to
> > google-appengi...@googlegroups.com<google-appengine%2Bunsubscrib e...@googlegroups.com>
> > .
> > > For more options, visit this group at
> >http://groups.google.com/group/google-appengine?hl=en.
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "Google App Engine" group.
> > To post to this group, send email to google-a...@googlegroups.com.
> > To unsubscribe from this group, send email to
> > google-appengi...@googlegroups.com<google-appengine%2Bunsubscrib e...@googlegroups.com>
johnP
the suggestion of lowering the severity of the log item from
'Critical' to 'Warning' or 'Info'.
Thanks!
johnP
On Apr 10, 7:18 am, notreadbyhumans <notreadbyhum...@gmail.com> wrote:
> Could someone offer a bit more detail regarding the meaning of this
> message. I have just started receiving it on what I thought was a
> fairly minor operation (calling ~1,500 records from the datastore and
> then looping through them to create a list based on various
> properties). I now get the following message each time this process
> runs:
>
> Exceededsoftmemorylimitwith 187.34 MB after servicing 1 requests
> total
>
> This seems very severe given the nature of the process that I am
> running (according to the dashboard my entire datastore is only 31
> MB!)
>
> On Mar 26, 3:36 am, Eli Jones <eli.jo...@gmail.com> wrote:
>
>
>
> > This is a nice addition.. I use a lot of per instance (or in clone cache or
> > whatever one may call it)memoryto cache datastore info (so.. the code
> > checks its local cache.. then memcache and then datastore).. but I've been
> > stuck sort of monitoring my tasks (I run looping tasks that access the same
> > data) to see if I can tell when they hit thememorylimitand do a cold
> > start.. and then reduce the size of the inmemorycache dict.
>
> > Now, I'm guessing this will explicitly tell me when it's hitting that
> >limit.. which will help a lot.
>
> > On Thu, Mar 25, 2010 at 1:51 PM, Jon McAlister <jon...@google.com> wrote:
> > > This is a new logging message we just added. You've most likely had
> > > clones exceed thesoftmemorylimitforever, but now we are letting
> > > you know via logs. These logs can of course be ignored if you like,
> > > since it doesn't effect the http response you served, but if you are
> > > concerned about performance or latency, then I would pay attention to
> > > these logs. The thing is, when a clone is killed after exceeding the
> > >softmemorylimit, it is dead and will receive no more requests. This
> > > means that the next request for your app will have to be a loading
> > > request, which depending on your app could mean it involves a much
> > > higher latency and cpu consumption. Reducing thememoryusage in your
> > > clones could then mean you have fewersoftmemorylimitterminations,
> > > and as such fewer loading requests.
>
> > > On Wed, Mar 24, 2010 at 9:19 PM, Blake <blake.commag...@gmail.com> wrote:
> > > > I also started seeing this just last night for the first time. I
> > > > haven't pushed any updates, so I don't believe it's an error on my
> > > > end.
> > > > Anyone else?
>
> > > > On Mar 23, 7:23 pm, Jackie Fei <3s4m...@gmail.com> wrote:
> > > >> Dear all,
>
> > > >> "Exceededsoftmemorylimitwith 187.727 MB after servicing 1 requests
peterk
it's ~187MB but can that be confirmed? Is it a fixed limit? It won't
be revised downward, will it?
Also, for those seeing these notices...just to clarify, they cause the
request to not be finished, right? It's not just a warning but
terminates the request?