Alaa Abd El Fattah schrieb:
> you can verify this using the memcached-tool command (if it's not
> installed on your system get it from
> http://code.sixapart.com/svn/memcached/trunk/server/scripts/memcached-tool
> it's a simple perl script)
>
> just run
> memcached-tool localhost stats
>
> the interesting values are get_misses and evictions but please send the
> whole output here.
>
>
I made some tests with the memcached-tool (see below). I can't find a
problem here, but always the first mainpage-call after a
refresh_stats-task costs about 2 minutes.
First I restarted the memcache-daemon:
#localhost:11211 Field Value
accepting_conns 1
bytes 0
bytes_read 7
bytes_written 0
cmd_flush 0
cmd_get 0
cmd_set 0
connection_structures 6
curr_connections 5
curr_items 0
evictions 0
get_hits 0
get_misses 0
limit_maxbytes 524288000
listen_disabled_num 0
pid 16510
pointer_size 64
rusage_system 0.000000
rusage_user 0.000000
threads 5
time 1262680944
total_connections 6
total_items 0
uptime 14
version 1.2.8
Then I run the refresh_stats-Task. After that I got this:
#localhost:11211 Field Value
accepting_conns 1
bytes 826655
bytes_read 2388370
bytes_written 429601
cmd_flush 0
cmd_get 3206
cmd_set 1783
connection_structures 6
curr_connections 5
curr_items 1685
evictions 0
get_hits 0
get_misses 3206
limit_maxbytes 524288000
listen_disabled_num 0
pid 16510
pointer_size 64
rusage_system 1.020000
rusage_user 0.310000
threads 5
time 1262681496
total_connections 8
total_items 1783
uptime 566
version 1.2.8
The first load of mainpage took about 2 minutes. After that I got this:
#localhost:11211 Field Value
accepting_conns 1
bytes 2045705
bytes_read 3653388
bytes_written 507863
cmd_flush 0
cmd_get 6678
cmd_set 5155
connection_structures 6
curr_connections 5
curr_items 4940
evictions 0
get_hits 100
get_misses 6578
limit_maxbytes 524288000
listen_disabled_num 0
pid 16510
pointer_size 64
rusage_system 1.220000
rusage_user 0.420000
threads 5
time 1262681647
total_connections 10
total_items 5155
uptime 717
version 1.2.8
Now I run my synchronisation-Task, which touches/rewrites nearly each
xliff-file. Afterwards I run the refresh_stats-Task again and got this:
#localhost:11211 Field Value
accepting_conns 1
bytes 584204
bytes_read 5808235
bytes_written 978912
cmd_flush 0
cmd_get 10064
cmd_set 6433
connection_structures 7
curr_connections 6
curr_items 1273
evictions 0
get_hits 231
get_misses 9833
limit_maxbytes 524288000
listen_disabled_num 0
pid 16510
pointer_size 64
rusage_system 2.390000
rusage_user 0.700000
threads 5
time 1262687590
total_connections 32
total_items 6433
uptime 6660
version 1.2.8
When I now try to open the pootle-mainpage it took about 18 minutes.
Than I got this:
#localhost:11211 Field Value
accepting_conns 1
bytes 1633582
bytes_read 7911596
bytes_written 1338432
cmd_flush 0
cmd_get 13392
cmd_set 9580
connection_structures 7
curr_connections 5
curr_items 4227
evictions 0
get_hits 417
get_misses 12975
limit_maxbytes 524288000
listen_disabled_num 0
pid 16510
pointer_size 64
rusage_system 3.460000
rusage_user 1.190000
threads 5
time 1262689669
total_connections 33
total_items 9580
uptime 8739
version 1.2.8
In a second step, I run the memcache-daemon in verbose-mode and made
some comparisons on the memcached.log-file. The refresh-task deletes the
":getquickstats" and the ":getcompletestats"-element of an xliff-file.
But it only adds a ":getcompletestats"-elements. I can't find any
adding-operation of a ":getquickstats"-element.
<31 delete /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats 0
> NOT FOUND /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats
>31 NOT_FOUND
<31 delete /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats 0
> FOUND KEY /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats
>31 DELETED
[...]
<31 get /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats
> NOT FOUND /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats
>31 END
<31 set /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats 1 2500000 164
> NOT FOUND /ar/f4pos_06/POS2/dataDictionary.xlf:getcompletestats
>31 STORED
The mainpage-call looks for a ":getquickstats"-element of the same file
only, which can't be found and is recalculated and stored in cache.
<31 get /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats
> NOT FOUND /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats
>31 END
<31 set /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats 1 2500000 261
> NOT FOUND /ar/f4pos_06/POS2/dataDictionary.xlf:getquickstats
>31 STORED
I think this is a bug in Pootle, isn't it?
greets
Johannes
------------------------------------------------------------------------------
This SF.Net email is sponsored by the Verizon Developer Community
Take advantage of Verizon's best-in-class app development support
A streamlined, 14 day to market process makes app distribution fast and easy
Join now and get one step closer to millions of Verizon customers
http://p.sf.net/sfu/verizon-dev2dev
_______________________________________________
Translate-pootle mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/translate-pootle
