Hi Piggers - Seeing an issue with a particular script where our job is taking 6hrs 42min to complete.  

syslogs are showing loads of these:
INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory handler called (Usage threshold exceeded) init = 5439488(5312K) used = 283443200(276800K) committed = 357957632(349568K) max = 357957632(349568K)
INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory handler called (Usage threshold exceeded) init = 5439488(5312K) used = 267128840(260868K) committed = 357957632(349568K) max = 357957632(349568K)
One iteresting thing is it's the map phase that is slow and one of the mappers is getting 8GB of input while the other 2000 or so mappers are getting MBs and hundreds of MBs of data.

Any where I can start looking?

Which version of Pig are you using?

-----Original Message-----
From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 06, 2010 10:29 AM
To: [EMAIL PROTECTED]
Subject: SpillableMemoryManager - low memory handler called

Hi Piggers - Seeing an issue with a particular script where our job is
taking 6hrs 42min to complete.  

syslogs are showing loads of these:
INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
handler called (Usage threshold exceeded) init = 5439488(5312K) used 283443200(276800K) committed = 357957632(349568K) max 357957632(349568K)
INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
handler called (Usage threshold exceeded) init = 5439488(5312K) used 267128840(260868K) committed = 357957632(349568K) max 357957632(349568K)
One iteresting thing is it's the map phase that is slow and one of the
mappers is getting 8GB of input while the other 2000 or so mappers are
getting MBs and hundreds of MBs of data.

Any where I can start looking?

0.6

Sent from my iPhone

On May 6, 2010, at 12:16 PM, "Olga Natkovich" <[EMAIL PROTECTED]>  
wrote:

> Which version of Pig are you using?
>
> -----Original Message-----
> From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, May 06, 2010 10:29 AM
> To: [EMAIL PROTECTED]
> Subject: SpillableMemoryManager - low memory handler called
>
> Hi Piggers - Seeing an issue with a particular script where our job is
> taking 6hrs 42min to complete.
>
> syslogs are showing loads of these:
> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
> handler called (Usage threshold exceeded) init = 5439488(5312K) used > 283443200(276800K) committed = 357957632(349568K) max > 357957632(349568K)
> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
> handler called (Usage threshold exceeded) init = 5439488(5312K) used > 267128840(260868K) committed = 357957632(349568K) max > 357957632(349568K)
> One iteresting thing is it's the map phase that is slow and one of the
> mappers is getting 8GB of input while the other 2000 or so mappers are
> getting MBs and hundreds of MBs of data.
>
> Any where I can start looking?
>
>

This is just a warning saying that your job is spilling to the disk.
Please, if you can, post a script that is causing this issue. In 0.6.0
we moved large chunk of the code away from using SpillableMemoryManager
but it is still used in some places. More changes are coming in 0.7.0 as
well.

Olga

-----Original Message-----
From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 06, 2010 11:31 AM
To: [EMAIL PROTECTED]
Subject: Re: SpillableMemoryManager - low memory handler called

0.6

Sent from my iPhone

On May 6, 2010, at 12:16 PM, "Olga Natkovich" <[EMAIL PROTECTED]>  
wrote:

> Which version of Pig are you using?
>
> -----Original Message-----
> From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, May 06, 2010 10:29 AM
> To: [EMAIL PROTECTED]
> Subject: SpillableMemoryManager - low memory handler called
>
> Hi Piggers - Seeing an issue with a particular script where our job is
> taking 6hrs 42min to complete.
>
> syslogs are showing loads of these:
> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
> handler called (Usage threshold exceeded) init = 5439488(5312K) used > 283443200(276800K) committed = 357957632(349568K) max > 357957632(349568K)
> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
> handler called (Usage threshold exceeded) init = 5439488(5312K) used > 267128840(260868K) committed = 357957632(349568K) max > 357957632(349568K)
> One iteresting thing is it's the map phase that is slow and one of the
> mappers is getting 8GB of input while the other 2000 or so mappers are
> getting MBs and hundreds of MBs of data.
>
> Any where I can start looking?
>
>

I was just looking at the SpillableMemoryManager code, and discovered
that it prints out this message even when it doesn't actually spill
anything, just when the GC gets called.
To reproduce, try using the TupleFactory to generate a few million
tuples. There's nothing to spill, and yet you will see the message
when GC kicks in.

But that's beside the point -- clearly something is wrong with the way
Corbin's data is getting distributed, if he gets gigs on one machine,
and megs on the rest. Please post the script :-).

-D

On Thu, May 6, 2010 at 11:36 AM, Olga Natkovich <[EMAIL PROTECTED]> wrote:
> This is just a warning saying that your job is spilling to the disk.
> Please, if you can, post a script that is causing this issue. In 0.6.0
> we moved large chunk of the code away from using SpillableMemoryManager
> but it is still used in some places. More changes are coming in 0.7.0 as
> well.
>
> Olga
>
> -----Original Message-----
> From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, May 06, 2010 11:31 AM
> To: [EMAIL PROTECTED]
> Subject: Re: SpillableMemoryManager - low memory handler called
>
> 0.6
>
> Sent from my iPhone
>
> On May 6, 2010, at 12:16 PM, "Olga Natkovich" <[EMAIL PROTECTED]>
> wrote:
>
>> Which version of Pig are you using?
>>
>> -----Original Message-----
>> From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
>> Sent: Thursday, May 06, 2010 10:29 AM
>> To: [EMAIL PROTECTED]
>> Subject: SpillableMemoryManager - low memory handler called
>>
>> Hi Piggers - Seeing an issue with a particular script where our job is
>> taking 6hrs 42min to complete.
>>
>> syslogs are showing loads of these:
>> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
>> handler called (Usage threshold exceeded) init = 5439488(5312K) used >> 283443200(276800K) committed = 357957632(349568K) max >> 357957632(349568K)
>> INFO : org.apache.pig.impl.util.SpillableMemoryManager - low memory
>> handler called (Usage threshold exceeded) init = 5439488(5312K) used >> 267128840(260868K) committed = 357957632(349568K) max >> 357957632(349568K)
>> One iteresting thing is it's the map phase that is slow and one of the
>> mappers is getting 8GB of input while the other 2000 or so mappers are
>> getting MBs and hundreds of MBs of data.
>>
>> Any where I can start looking?
>>
>>
>

-----Original Message-----
From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 06, 2010 11:57 AM
To: Olga Natkovich
Subject: Re: SpillableMemoryManager - low memory handler called

I have attached the script... please let me know if you have more
questions.  

Looks like attachments are not coming through. Here is the script from
Corbin inline.

One thing you might want to try is to switch your cogroups to skewed
join and see if that solves the issue:

http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins

Olga

--------------------------------------------topurl.pig------------------
-------------------------------------------
set job.name 'Generate topurl reports for $out_file1'

%default dir_prefix '../..'
%default storage 'BinStorage()'
%default tynt_udfs 'tynt-udfs.jar'
%default topN '20'
/* default to 30 days time period so that alltime report will get
14*30=420 min page views*/
%default timeperiod '30'
%default min_page_views_per_day '14'
 
register $dir_prefix/udfs/target/$tynt_udfs
register $dir_prefix/udfs/lib/piggybank.jar

---------------------summarize address bar
stats-----------------------------------
addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
(site:chararray, url:chararray, guid:chararray, cnt:long);
grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
SUM(addbar_stats.cnt) AS addbarvisits;
STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
$storage;

grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
PARALLEL 180;
addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
$storage;

----------------------calculate
ratio------------------------------------------
clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
(site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
uwcnt:long);
viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
(site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);

light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url, cnt;
light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site, url,
title, cnt, etcnt;

light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url != '-';

--light_addbarstatsbyurl = FOREACH addbar_stats_by_url GENERATE site,
url, addbarvisits;
--joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
INNER, light_clickstatsbyurl BY (site, url) OUTER,
light_addbarstatsbyurl BY (site, url) OUTER;
--flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
--
(IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
clickcnt,
--
(IsEmpty(light_addbarstatsbyurl)?0:MAX(light_addbarstatsbyurl.addbarvisi
ts)) as addbarcnt;

joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
INNER, light_clickstatsbyurl BY (site, url) OUTER;
flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
 
(IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
clickcnt;

ratio_by_url = FOREACH flattened_stats_for_ratio
                      {
                        generated_traffic = clickcnt+etcnt;
                        total_traffic = cnt;
                        ti ((float)(generated_traffic))/((float)total_traffic);
                        GENERATE site, url, title, ((ti>1)?(-ti):ti) AS
ratio, generated_traffic AS gviews, total_traffic AS views;
                      }

------------------------combined with
#copies----------------------------------------
copystatsbyurl = LOAD '$in_file1/copystatsbyurl' USING $storage AS
(site:chararray, url:chararray, lcnt:long, scnt:long, icnt:long,
acnt:long);
light_copystatsbyurl = FOREACH copystatsbyurl GENERATE site, url,
lcnt+scnt+icnt AS cnt;

all_stats_by_url = COGROUP ratio_by_url BY (site, url) INNER,
light_copystatsbyurl BY (site, url) OUTER PARALLEL 62;
all_urls = FOREACH all_stats_by_url GENERATE FLATTEN(ratio_by_url) AS
(site, url, title, ratio, gviews, views),
(IsEmpty(light_copystatsbyurl)?0:MAX(light_copystatsbyurl.cnt)) as
copies;

grouped_urls_by_site = GROUP all_urls BY site;

top_ratios = FOREACH grouped_urls_by_site
                  {
                    filtered_by_minpageviews = FILTER all_urls BY views
                    order_by_ratio = ORDER filtered_by_minpageviews BY
ratio DESC;
                    top_by_ratio = LIMIT order_by_ratio $topN;
                    GENERATE group AS site, top_by_ratio.(url, title,
ratio, gviews, views, copies) AS tops;                  
                  }

top_gviews = FOREACH grouped_urls_by_site
                  {
                    order_by_gviews = ORDER all_urls BY gviews DESC;
                    top_by_gviews = LIMIT order_by_gviews $topN;
                    GENERATE group AS site, top_by_gviews.(url, title,
ratio, gviews, views, copies) AS tops;                  
                  }

top_views = FOREACH grouped_urls_by_site
                  {
                    order_by_views = ORDER all_urls BY views DESC;
                    top_by_views = LIMIT order_by_views $topN;
                    GENERATE group AS site, top_by_views.(url, title,
ratio, gviews, views, copies) AS tops;                  
                  }

top_copies = FOREACH grouped_urls_by_site
                  {
                    order_by_copies = ORDER all_urls BY copies DESC;
                    top_by_copies = LIMIT order_by_copies $topN;
                    GENERATE group AS site, top_by_copies.(url, title,
ratio, gviews, views, copies) AS tops;                  
                  }

grouped_tops = JOIN top_ratios BY site, top_gviews BY site, top_views BY
site, top_copies BY site;

top_urls = FOREACH grouped_t

Wondering if when we do a group like this:

grouped_urls_by_site = GROUP all_urls BY site;

if certain site has a lot of urls would they all have to be processed by the same mapper (e.g. a single key?)  Could this account for why we have 8GB in one map and not many in others?

On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:

> Looks like attachments are not coming through. Here is the script from
> Corbin inline.
>
> One thing you might want to try is to switch your cogroups to skewed
> join and see if that solves the issue:
>
> http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>
> Olga
>
> --------------------------------------------topurl.pig------------------
> -------------------------------------------
> set job.name 'Generate topurl reports for $out_file1'
>
> %default dir_prefix '../..'
> %default storage 'BinStorage()'
> %default tynt_udfs 'tynt-udfs.jar'
> %default topN '20'
> /* default to 30 days time period so that alltime report will get
> 14*30=420 min page views*/
> %default timeperiod '30'
> %default min_page_views_per_day '14'
>
> register $dir_prefix/udfs/target/$tynt_udfs
> register $dir_prefix/udfs/lib/piggybank.jar
>
> ---------------------summarize address bar
> stats-----------------------------------
> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
> (site:chararray, url:chararray, guid:chararray, cnt:long);
> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
> SUM(addbar_stats.cnt) AS addbarvisits;
> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
> $storage;
>
> grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
> PARALLEL 180;
> addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
> group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
> SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
> STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
> $storage;
>
> ----------------------calculate
> ratio------------------------------------------
> clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
> (site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
> pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
> uwcnt:long);
> viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
> (site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
> et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);
>
> light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url, cnt;
> light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site, url,
> title, cnt, etcnt;
>
> light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url != '-';
>
> --light_addbarstatsbyurl = FOREACH addbar_stats_by_url GENERATE site,
> url, addbarvisits;
> --joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
> INNER, light_clickstatsbyurl BY (site, url) OUTER,
> light_addbarstatsbyurl BY (site, url) OUTER;
> --flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
> FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
> --
> (IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
> clickcnt,
> --
> (IsEmpty(light_addbarstatsbyurl)?0:MAX(light_addbarstatsbyurl.addbarvisi
> ts)) as addbarcnt;
>
> joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
> INNER, light_clickstatsbyurl BY (site, url) OUTER;
> flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
> FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
>
> (IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
> clickcnt;
>
> ratio_by_url = FOREACH flattened_stats_for_ratio
>                      {
>                        generated_traffic = clickcnt+etcnt;
>                        total_traffic = cnt;
>                        ti > ((float)(generated_traffic))/((float)total_traffic);

Groups are processed on the reduce side with combiner pre-aggregating on
the map side. A single map by default gets a fixed chunk of data and not
the entire key. The only way I can see a huge map happening is if you
have one really huge record somewhere.

Olga

-----Original Message-----
From: Corbin Hoenes [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 06, 2010 2:31 PM
To: [EMAIL PROTECTED]
Subject: Re: SpillableMemoryManager - low memory handler called

Wondering if when we do a group like this:

grouped_urls_by_site = GROUP all_urls BY site;

if certain site has a lot of urls would they all have to be processed by
the same mapper (e.g. a single key?)  Could this account for why we have
8GB in one map and not many in others?

On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:

> Looks like attachments are not coming through. Here is the script from
> Corbin inline.
>
> One thing you might want to try is to switch your cogroups to skewed
> join and see if that solves the issue:
>
>
http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>
> Olga
>
>
--------------------------------------------topurl.pig------------------
> -------------------------------------------
> set job.name 'Generate topurl reports for $out_file1'
>
> %default dir_prefix '../..'
> %default storage 'BinStorage()'
> %default tynt_udfs 'tynt-udfs.jar'
> %default topN '20'
> /* default to 30 days time period so that alltime report will get
> 14*30=420 min page views*/
> %default timeperiod '30'
> %default min_page_views_per_day '14'
>
> register $dir_prefix/udfs/target/$tynt_udfs
> register $dir_prefix/udfs/lib/piggybank.jar
>
> ---------------------summarize address bar
> stats-----------------------------------
> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
> (site:chararray, url:chararray, guid:chararray, cnt:long);
> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL
180;
> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
> SUM(addbar_stats.cnt) AS addbarvisits;
> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
> $storage;
>
> grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
> PARALLEL 180;
> addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
> group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
> SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
> STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
> $storage;
>
> ----------------------calculate
> ratio------------------------------------------
> clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
> (site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
> pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
> uwcnt:long);
> viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
> (site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
> et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);
>
> light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url,
cnt;
> light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site,
url,
> title, cnt, etcnt;
>
> light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url !'-';
>
> --light_addbarstatsbyurl = FOREACH addbar_stats_by_url GENERATE site,
> url, addbarvisits;
> --joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
> INNER, light_clickstatsbyurl BY (site, url) OUTER,
> light_addbarstatsbyurl BY (site, url) OUTER;
> --flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
> FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
> --
> (IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
> clickcnt,
> --
>
(IsEmpty(light_addbarstatsbyurl)?0:MAX(light_addbarstatsbyurl.addbarvisi
> ts)) as addbarcnt;
>
> joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
> INNER, light_clickstatsbyurl BY (site, url) OUTER;
BY

0.6.0
SpillableMemoryManager
as
is
the
are

Okay we did some filtering on the all_urls in the nested FOREACH and that seems to fix the performance issue.  There are still some mappers that get 8GB of data but the job went down to 2 hours.  

From Dimtry's reply sounds like the low memory handler output in the logs is misleading.

On May 6, 2010, at 3:30 PM, Corbin Hoenes wrote:

> Wondering if when we do a group like this:
>
> grouped_urls_by_site = GROUP all_urls BY site;
>
> if certain site has a lot of urls would they all have to be processed by the same mapper (e.g. a single key?)  Could this account for why we have 8GB in one map and not many in others?
>
> On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:
>
>> Looks like attachments are not coming through. Here is the script from
>> Corbin inline.
>>
>> One thing you might want to try is to switch your cogroups to skewed
>> join and see if that solves the issue:
>>
>> http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>>
>> Olga
>>
>> --------------------------------------------topurl.pig------------------
>> -------------------------------------------
>> set job.name 'Generate topurl reports for $out_file1'
>>
>> %default dir_prefix '../..'
>> %default storage 'BinStorage()'
>> %default tynt_udfs 'tynt-udfs.jar'
>> %default topN '20'
>> /* default to 30 days time period so that alltime report will get
>> 14*30=420 min page views*/
>> %default timeperiod '30'
>> %default min_page_views_per_day '14'
>>
>> register $dir_prefix/udfs/target/$tynt_udfs
>> register $dir_prefix/udfs/lib/piggybank.jar
>>
>> ---------------------summarize address bar
>> stats-----------------------------------
>> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
>> (site:chararray, url:chararray, guid:chararray, cnt:long);
>> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
>> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
>> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
>> SUM(addbar_stats.cnt) AS addbarvisits;
>> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
>> $storage;
>>
>> grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
>> PARALLEL 180;
>> addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
>> group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
>> SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
>> STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
>> $storage;
>>
>> ----------------------calculate
>> ratio------------------------------------------
>> clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
>> (site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
>> pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
>> uwcnt:long);
>> viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
>> (site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
>> et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);
>>
>> light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url, cnt;
>> light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site, url,
>> title, cnt, etcnt;
>>
>> light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url != '-';
>>
>> --light_addbarstatsbyurl = FOREACH addbar_stats_by_url GENERATE site,
>> url, addbarvisits;
>> --joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
>> INNER, light_clickstatsbyurl BY (site, url) OUTER,
>> light_addbarstatsbyurl BY (site, url) OUTER;
>> --flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
>> FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
>> --
>> (IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as
>> clickcnt,
>> --
>> (IsEmpty(light_addbarstatsbyurl)?0:MAX(light_addbarstatsbyurl.addbarvisi
>> ts)) as addbarcnt;
>>
>> joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
>> INNER, light_clickstatsbyurl BY (site, url) OUTER;

Hi,
   Do you know which snippet in the script is causing the issue ?
There are multiple MR jobs which will be executed, what is causing the
exact issue ?
Map side spills is strange - are you sure it is not in the reducer ?

If it really is in the map side, I guess it is pointing to the case
where a single record (which is being read) is so large that it is
triggering spills : this could theoretically be possible if you store
and then load it again - where reducers are set to higher memory than
mappers. Though I did not see something like this happening here ...

Regards,
Mridul
On Friday 07 May 2010 03:00 AM, Corbin Hoenes wrote:
> Wondering if when we do a group like this:
>
> grouped_urls_by_site = GROUP all_urls BY site;
>
> if certain site has a lot of urls would they all have to be processed by the same mapper (e.g. a single key?)  Could this account for why we have 8GB in one map and not many in others?
>
> On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:
>
>> Looks like attachments are not coming through. Here is the script from
>> Corbin inline.
>>
>> One thing you might want to try is to switch your cogroups to skewed
>> join and see if that solves the issue:
>>
>> http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>>
>> Olga
>>
>> --------------------------------------------topurl.pig------------------
>> -------------------------------------------
>> set job.name 'Generate topurl reports for $out_file1'
>>
>> %default dir_prefix '../..'
>> %default storage 'BinStorage()'
>> %default tynt_udfs 'tynt-udfs.jar'
>> %default topN '20'
>> /* default to 30 days time period so that alltime report will get
>> 14*30=420 min page views*/
>> %default timeperiod '30'
>> %default min_page_views_per_day '14'
>>
>> register $dir_prefix/udfs/target/$tynt_udfs
>> register $dir_prefix/udfs/lib/piggybank.jar
>>
>> ---------------------summarize address bar
>> stats-----------------------------------
>> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
>> (site:chararray, url:chararray, guid:chararray, cnt:long);
>> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
>> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
>> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
>> SUM(addbar_stats.cnt) AS addbarvisits;
>> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
>> $storage;
>>
>> grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
>> PARALLEL 180;
>> addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
>> group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
>> SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
>> STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
>> $storage;
>>
>> ----------------------calculate
>> ratio------------------------------------------
>> clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
>> (site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
>> pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
>> uwcnt:long);
>> viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
>> (site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
>> et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);
>>
>> light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url, cnt;
>> light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site, url,
>> title, cnt, etcnt;
>>
>> light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url != '-';
>>
>> --light_addbarstatsbyurl = FOREACH addbar_stats_by_url GENERATE site,
>> url, addbarvisits;
>> --joined_stats_for_ratio = COGROUP light_viewstatsbyurl BY (site, url)
>> INNER, light_clickstatsbyurl BY (site, url) OUTER,
>> light_addbarstatsbyurl BY (site, url) OUTER;
>> --flattened_stats_for_ratio = FOREACH joined_stats_for_ratio GENERATE
>> FLATTEN(light_viewstatsbyurl) AS (site, url, title, cnt, etcnt),
>> --
>> (IsEmpty(light_clickstatsbyurl)?0:MAX(light_clickstatsbyurl.cnt)) as

Mridul,
like I said earlier -- unfortunately at the moment those log lines
only mean that the Java garbage collector is invoked, not that
anything is actually spilled. This gets printed out even if there are
no spillable objects the Manager is aware of. An 8G map will certainly
trigger the GC.

On Fri, May 7, 2010 at 2:44 PM, Mridul Muralidharan
<[EMAIL PROTECTED]> wrote:
>
> Hi,
>
>
>  Do you know which snippet in the script is causing the issue ?
> There are multiple MR jobs which will be executed, what is causing the exact
> issue ?
>
>
> Map side spills is strange - are you sure it is not in the reducer ?
>
> If it really is in the map side, I guess it is pointing to the case where a
> single record (which is being read) is so large that it is triggering spills
> : this could theoretically be possible if you store and then load it again -
> where reducers are set to higher memory than mappers. Though I did not see
> something like this happening here ...
>
>
>
> Regards,
> Mridul
>
>
>
>
> On Friday 07 May 2010 03:00 AM, Corbin Hoenes wrote:
>>
>> Wondering if when we do a group like this:
>>
>> grouped_urls_by_site = GROUP all_urls BY site;
>>
>> if certain site has a lot of urls would they all have to be processed by
>> the same mapper (e.g. a single key?)  Could this account for why we have 8GB
>> in one map and not many in others?
>>
>> On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:
>>
>>> Looks like attachments are not coming through. Here is the script from
>>> Corbin inline.
>>>
>>> One thing you might want to try is to switch your cogroups to skewed
>>> join and see if that solves the issue:
>>>
>>> http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>>>
>>> Olga
>>>
>>> --------------------------------------------topurl.pig------------------
>>> -------------------------------------------
>>> set job.name 'Generate topurl reports for $out_file1'
>>>
>>> %default dir_prefix '../..'
>>> %default storage 'BinStorage()'
>>> %default tynt_udfs 'tynt-udfs.jar'
>>> %default topN '20'
>>> /* default to 30 days time period so that alltime report will get
>>> 14*30=420 min page views*/
>>> %default timeperiod '30'
>>> %default min_page_views_per_day '14'
>>>
>>> register $dir_prefix/udfs/target/$tynt_udfs
>>> register $dir_prefix/udfs/lib/piggybank.jar
>>>
>>> ---------------------summarize address bar
>>> stats-----------------------------------
>>> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
>>> (site:chararray, url:chararray, guid:chararray, cnt:long);
>>> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
>>> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
>>> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
>>> SUM(addbar_stats.cnt) AS addbarvisits;
>>> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
>>> $storage;
>>>
>>> grouped_addbar_stats_by_site = GROUP addbar_stats_by_url BY site
>>> PARALLEL 180;
>>> addbar_stats_by_site = FOREACH grouped_addbar_stats_by_site GENERATE
>>> group AS site, SUM(addbar_stats_by_url.addbarcnt) AS addbarcnt,
>>> SUM(addbar_stats_by_url.addbarvisits) AS addbarvisits;
>>> STORE addbar_stats_by_site INTO '$out_file1/addbarstatsbysite' USING
>>> $storage;
>>>
>>> ----------------------calculate
>>> ratio------------------------------------------
>>> clickstatsbyurl = LOAD '$in_file1/clickstatsbyurl' USING $storage AS
>>> (site:chararray, url:chararray, cnt:long, tracecnt:long, tcnt:long,
>>> pcnt:long, wcnt:long, utracecnt:long, utcnt:long, upcnt:long,
>>> uwcnt:long);
>>> viewstatsbyurl = LOAD '$in_file1/viewstatsbyurl' USING $storage AS
>>> (site:chararray, url:chararray, title:chararray, cnt:long, etcnt:long,
>>> et1cnt:long, et2cnt:long, et3cnt:long, et6cnt:long, et7cnt:long);
>>>
>>> light_clickstatsbyurl = FOREACH clickstatsbyurl GENERATE site, url, cnt;
>>> light_viewstatsbyurl_noisy = FOREACH viewstatsbyurl GENERATE site, url,
>>> title, cnt, etcnt;
>>>
>>> light_viewstatsbyurl = FILTER light_viewstatsbyurl_noisy BY url != '-';

The main memory intensive aspect of pig are -
a) bags with large number of tuples,
b) udf consuming memory,
c) very very large tuples.
d) bugs/memory leaks in pig/udf.

(a) is the most common case - particularly because of joins, cogroups,
pig internally using bag's in its pipeline, etc.

There is nothing much we can do about (c) - since it is a practical
limitation which affects direct mapreduce too.

(b) is rare, though possible - and requires user intervention : iirc
none of the stock pig udf's exhibit this issue.

(d) is a remote possibility, but hopefully it is not the case !

It is unusual for mapper to run out of memory except for (b), (c).

Unless ...
Now, if the reducer could generate this - but the mapper could not
consume the output of the reducer, it typically points to very large
bags such that it fits into reducer's memory, but since mapper has less
memory than reducer (usual hadoop config atleast in y!), it fails there.

Regards,
Mridul
On Saturday 08 May 2010 03:26 AM, Dmitriy Ryaboy wrote:
> Mridul,
> like I said earlier -- unfortunately at the moment those log lines
> only mean that the Java garbage collector is invoked, not that
> anything is actually spilled. This gets printed out even if there are
> no spillable objects the Manager is aware of. An 8G map will certainly
> trigger the GC.
>
> On Fri, May 7, 2010 at 2:44 PM, Mridul Muralidharan
> <[EMAIL PROTECTED]>  wrote:
>>
>> Hi,
>>
>>
>>   Do you know which snippet in the script is causing the issue ?
>> There are multiple MR jobs which will be executed, what is causing the exact
>> issue ?
>>
>>
>> Map side spills is strange - are you sure it is not in the reducer ?
>>
>> If it really is in the map side, I guess it is pointing to the case where a
>> single record (which is being read) is so large that it is triggering spills
>> : this could theoretically be possible if you store and then load it again -
>> where reducers are set to higher memory than mappers. Though I did not see
>> something like this happening here ...
>>
>>
>>
>> Regards,
>> Mridul
>>
>>
>>
>>
>> On Friday 07 May 2010 03:00 AM, Corbin Hoenes wrote:
>>>
>>> Wondering if when we do a group like this:
>>>
>>> grouped_urls_by_site = GROUP all_urls BY site;
>>>
>>> if certain site has a lot of urls would they all have to be processed by
>>> the same mapper (e.g. a single key?)  Could this account for why we have 8GB
>>> in one map and not many in others?
>>>
>>> On May 6, 2010, at 3:24 PM, Olga Natkovich wrote:
>>>
>>>> Looks like attachments are not coming through. Here is the script from
>>>> Corbin inline.
>>>>
>>>> One thing you might want to try is to switch your cogroups to skewed
>>>> join and see if that solves the issue:
>>>>
>>>> http://hadoop.apache.org/pig/docs/r0.6.0/piglatin_ref1.html#Skewed+Joins
>>>>
>>>> Olga
>>>>
>>>> --------------------------------------------topurl.pig------------------
>>>> -------------------------------------------
>>>> set job.name 'Generate topurl reports for $out_file1'
>>>>
>>>> %default dir_prefix '../..'
>>>> %default storage 'BinStorage()'
>>>> %default tynt_udfs 'tynt-udfs.jar'
>>>> %default topN '20'
>>>> /* default to 30 days time period so that alltime report will get
>>>> 14*30=420 min page views*/
>>>> %default timeperiod '30'
>>>> %default min_page_views_per_day '14'
>>>>
>>>> register $dir_prefix/udfs/target/$tynt_udfs
>>>> register $dir_prefix/udfs/lib/piggybank.jar
>>>>
>>>> ---------------------summarize address bar
>>>> stats-----------------------------------
>>>> addbar_stats = LOAD '$in_file1/addbarstats' USING $storage AS
>>>> (site:chararray, url:chararray, guid:chararray, cnt:long);
>>>> grouped_addbar_by_url = GROUP addbar_stats BY (site, url) PARALLEL 180;
>>>> addbar_stats_by_url = FOREACH grouped_addbar_by_url GENERATE
>>>> FLATTEN(group) AS (site, url), COUNT(addbar_stats) AS addbarcnt,
>>>> SUM(addbar_stats.cnt) AS addbarvisits;
>>>> STORE addbar_stats_by_url INTO '$out_file1/addbarstatsbyurl' USING
>>>

sorry to resurrect a old thread. But I am having the same problem here for
Pig 0.7

011-10-19 13:03:22,895 INFO org.apache.pig.impl.util.SpillableMemoryManager:
low memory handler called (Usage threshold exceeded) init 263323648(257152K) used = 502624976(490844K) committed = 617086976(602624K)
max = 715849728(699072K) 2011-10-19 13:03:58,776 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 494572568(482981K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:04:01,248 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Usage threshold exceeded) init = 263323648(257152K) used 501339160(489589K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:05:01,250 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 589147376(575339K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:05:42,181 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 487209296(475790K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:06:07,300 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Usage threshold exceeded) init = 263323648(257152K) used 501963568(490198K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:07:11,850 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 575026232(561549K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:07:53,782 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 505687176(493835K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:09:02,451 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 571977464(558571K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:09:44,027 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 509238248(497302K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:10:55,596 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 561933048(548762K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:11:40,931 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 517425112(505297K) committed = 715849728(699072K) max = 715849728(699072K)
2011-10-19 13:12:48,891 INFO
org.apache.pig.impl.util.SpillableMemoryManager: low memory handler called
(Collection threshold exceeded) init = 263323648(257152K) used 558586544(545494K) committed = 715849728(699072K) max = 715849728(699072K)

The job took a while to finish and not sure what causes it, it only happens
in the reduce phase and i am wondering how to increase the max memory to be
used by pig. In my mapreduce.reduce.java.opts, I have set the value to 1GB
but it seems the max is only 700MB for whatever reason.

Thanks,

Felix