[FUG-BR] OT: Squid muito lento....

Enio Marconcini # www.Enio.Pro.Br eniorm em gmail.com
Quinta Maio 20 10:01:40 BRT 2010


2010/5/20 Ricardo Nascimento Ferreira <ricardonf em gmail.com>

> Enio,
> o cache.log não apresentou alguma mensagem interessante ?
>
> Ricardo SilverDragon
>
> 2010/5/20 Enio Marconcini # www.Enio.Pro.Br <eniorm em gmail.com>
>
> > amigos,
> > tenho um Squid 2.X (na realidade é o Lusca) rodando num servidor FreeBSD,
> > com 1.5G de ram e processador Pentim 4 Dual Core
> >
> > de uns tempos pra cá ele ficou muito lento, até achei que era o antivirus
> > integrado com o HAVP, mas o desativei, e mesmo assim o squid continua
> > lento...
> >
> >
> >
> > olhando o top, os processos do squid aparecem assim:
> >
> > last pid: 24327;  load averages:  0.01,  0.04,
> > 0.01                                                up 46+18:24:22
> >  09:24:01
> > 257 processes: 1 running, 256 sleeping
> > CPU:  0.0% user,  0.0% nice,  2.6% system,  0.4% interrupt, 97.0% idle
> > Mem: 400M Active, 469M Inact, 178M Wired, 3204K Cache, 112M Buf, 433M
> Free
> > Swap: 2048M Total, 536K Used, 2047M Free
> >
> >  PID USERNAME   THR PRI NICE   SIZE    RES STATE   C   TIME   WCPU
> COMMAND
> > 24285 squid        1  44    0  3444K  1116K select  0   0:00  0.00%
> pinger
> > 24259 squid        1   4    0  3508K  1072K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24264 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24273 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24266 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24262 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24274 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24260 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24267 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24269 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24270 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24268 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24278 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24271 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24277 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24265 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24272 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24275 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24261 squid        1   4    0  3508K   912K sbwait  1   0:00  0.00%
> > ncsa_auth
> > 24263 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24276 squid        1   4    0  3508K   912K sbwait  0   0:00  0.00%
> > ncsa_auth
> > 24279 squid        1  -8    0  1404K   500K piperd  0   0:00  0.00%
> unlinkd
> > 24256 squid        1   8    0 12724K  2448K wait    1   0:00  0.00% squid
> > 24258 squid       41  44    0 58840K 41376K ucond   1   0:00  0.00% squid
> >
> > pensei ser gargalo no disco, mas olhando com o gstat aparentemente o
> disco
> > está folgado,
> >
> > pra ter idéia ta lento até pra abrir a página do google,
> > se eu natear meu IP e passar direto por fora do squid, vai uma beleza...
> >
> > no squid.conf alguns ajustes estão assim
> >
> > # memoria
> > cache_mem                       512 MB
> > cache_swap_low                  90
> > cache_swap_high                 95
> > maximum_object_size             100 MB
> > minimum_object_size             0 KB
> > maximum_object_size_in_memory   16 KB
> >
> > auth_param                      basic program
> > /usr/local/libexec/squid/ncsa_auth /usr/local/etc/squid/passwd_squid
> > auth_param                      basic children 20
> > auth_param                      basic realm Para continuar, informe seu
> > Login e Senha de acesso a INTERNET. Lembre-se, todos
> > os sites acessados serao monitorados.
> >
> >
> > mas não consigo identificar o problema da lentidão,
> >
> > as partições cache estão sendo consumidas 20% somente da capacidade....
> >
> > qualquer dica será bem vinda!
> >
> > --
> > ENIO RODRIGO MARCONCINI
> > gtalk: eniorm em gmail.com
> > skype: eniorm
> > msn: /dev/null
> >
> > .: FreeBSD -:- OpenBSD -:-Slackware Linux :.
> > Have trouble with Windows - reboot!
> > Have trouble with Unix - be root!
> > -------------------------
> > Histórico: http://www.fug.com.br/historico/html/freebsd/
> > Sair da lista: https://www.fug.com.br/mailman/listinfo/freebsd
> >
>
>
>
> --
> Ricardo Nascimento Ferreira
> Analista de Segurança da Informação
> CISSP-DF Grupo de Estudos
> Modulo Certified Security Officer
> Solaris Certified System Administrator
> -------------------------
> Histórico: http://www.fug.com.br/historico/html/freebsd/
> Sair da lista: https://www.fug.com.br/mailman/listinfo/freebsd
>


Opa Ricardo,

apresentou sim, esqueci de colar no email anterior,

veja

2010/05/19 20:59:01| aioSync: flushing pending I/O operations
2010/05/19 20:59:01| aioSync: done
2010/05/19 20:59:01| aioSync: flushing pending I/O operations
2010/05/19 20:59:01| aioSync: done
2010/05/19 20:59:01| aioSync: flushing pending I/O operations
2010/05/19 20:59:01| aioSync: done
2010/05/19 20:59:01| aioSync: flushing pending I/O operations
2010/05/19 20:59:01| aioSync: done
2010/05/19 20:59:01| aioSync: flushing pending I/O operations
2010/05/19 20:59:01| aioSync: done
(...)
2010/05/20 07:01:14| logfileClose: closing log
/usr/local/squid/logs/netdb.state
2010/05/20 07:01:14| NETDB state saved; 788 entries, 6 msec
2010/05/20 08:10:48| logfileOpen: opening log
/usr/local/squid/logs/netdb.state
2010/05/20 08:10:48| logfileClose: closing log
/usr/local/squid/logs/netdb.state
2010/05/20 08:10:48| NETDB state saved; 794 entries, 5 msec
2010/05/20 08:56:59| ftpTimeout: timeout in SENT_PASV state
2010/05/20 08:58:26| clientProcessExpired: collapsed request STALE!
2010/05/20 09:08:52| ftpReadControlReply: read error: (60) Operation timed
out
2010/05/20 09:13:57| squidaio_queue_request: WARNING - Queue congestion
2010/05/20 09:33:50| Squid Cache (Version LUSCA_HEAD): Exiting normally.
2010/05/20 09:33:51| Starting Squid Cache version LUSCA_HEAD for
i386-portbld-freebsd7.3...
2010/05/20 09:33:51| Process ID 24410
2010/05/20 09:33:51| With 11095 file descriptors available
2010/05/20 09:33:51| Using kqueue for the IO loop
2010/05/20 09:33:51| Adding nameserver 127.0.0.1 from squid.conf
2010/05/20 09:33:51| DNS Socket created at 0.0.0.0, port 49836, FD 6
2010/05/20 09:33:51| helperOpenServers: Starting 20 'ncsa_auth' processes
2010/05/20 09:33:51| Unlinkd pipe opened on FD 30
2010/05/20 09:33:51| Swap maxSize 40960000 + 716800 KB, estimated 3205907
objects
2010/05/20 09:33:51| Target number of buckets: 160295
2010/05/20 09:33:51| Using 262144 Store buckets
2010/05/20 09:33:51| Max Mem  size: 716800 KB
2010/05/20 09:33:51| Max Swap size: 40960000 KB
2010/05/20 09:33:51| Store logging disabled
2010/05/20 09:33:51| Rebuilding storage in /squid/cache1 (CLEAN)
2010/05/20 09:33:51| Rebuilding storage in /squid/cache2 (CLEAN)
2010/05/20 09:33:51| Rebuilding storage in /squid/cache3 (CLEAN)
2010/05/20 09:33:51| Rebuilding storage in /squid/cache4 (CLEAN)
2010/05/20 09:33:51| Rebuilding storage in /squid/cache5 (CLEAN)
2010/05/20 09:33:51| Using Least Load store dir selection
2010/05/20 09:33:51| Current Directory is /usr/local/squid/logs
2010/05/20 09:33:51| Loaded Icons.
2010/05/20 09:33:51| Accepting proxy HTTP connections at 127.0.0.1, port
8080, FD 41.
2010/05/20 09:33:51| Accepting proxy HTTP connections at 192.168.0.13, port
8080, FD 42.
2010/05/20 09:33:51| Accepting SNMP messages on port 3401, FD 43.
2010/05/20 09:33:51| Pinger socket opened on FD 45
2010/05/20 09:33:51| NETDB state reloaded; 794 entries, 19 msec
2010/05/20 09:33:51| Ready to serve requests.
2010/05/20 09:33:51| pingerOpen: ICMP socket opened
2010/05/20 09:33:53| pingerReadRequest: socket 0: read() failed; errno 61
2010/05/20 09:33:53| pingerReadRequest: socket 0: read() failed; errno 61
2010/05/20 09:33:53| Pinger exiting.
2010/05/20 09:33:53| Pinger exiting.
2010/05/20 09:33:54| Store rebuilding is 100.0% complete
2010/05/20 09:33:54|   /squid/cache1: completed rebuild
2010/05/20 09:33:54| Done scanning /squid/cache1 (56720 entries)
2010/05/20 09:33:54|   /squid/cache3: completed rebuild
2010/05/20 09:33:54| Done scanning /squid/cache3 (61166 entries)
2010/05/20 09:33:54|   /squid/cache4: completed rebuild
2010/05/20 09:33:54| Done scanning /squid/cache4 (75144 entries)
2010/05/20 09:33:55|   /squid/cache5: completed rebuild
2010/05/20 09:33:55| Done scanning /squid/cache5 (76466 entries)
2010/05/20 09:33:55|   /squid/cache2: completed rebuild
2010/05/20 09:33:55| Done scanning /squid/cache2 (80039 entries)
2010/05/20 09:33:55| Finished rebuilding storage from disk.
2010/05/20 09:33:55|    349535 Entries scanned
2010/05/20 09:33:55|         0 Invalid entries.
2010/05/20 09:33:55|         0 With invalid flags.
2010/05/20 09:33:55|    324295 Objects loaded.
2010/05/20 09:33:55|         0 Objects expired.
2010/05/20 09:33:55|         0 Objects cancelled.
2010/05/20 09:33:55|         0 Duplicate URLs purged.
2010/05/20 09:33:55|     25240 Swapfile clashes avoided.
2010/05/20 09:33:55|   Took 3.6 seconds (90854.0 objects/sec).
2010/05/20 09:33:55| Beginning Validation Procedure
2010/05/20 09:33:55|    262144 Entries Validated so far.
2010/05/20 09:33:55|   Completed Validation Procedure
2010/05/20 09:33:55|   Validated 324295 Entries
2010/05/20 09:33:55|   store_swap_size = 10355928k
2010/05/20 09:33:55| storeLateRelease: released 0 objects
2010/05/20 09:43:54| squidaio_queue_request: WARNING - Queue congestion
2010/05/20 09:47:22| clientProcessExpired: collapsed request STALE!
2010/05/20 09:50:30| clientProcessExpired: collapsed request STALE!
2010/05/20 09:50:37| clientProcessExpired: collapsed request STALE!
2010/05/20 09:50:37| clientProcessExpired: collapsed request STALE!
2010/05/20 09:51:30| clientProcessExpired: collapsed request STALE!
2010/05/20 09:51:31| clientProcessExpired: collapsed request STALE!
2010/05/20 09:51:31| clientProcessExpired: collapsed request STALE!


Thiago, não uso nenhuma blacklists


-- 
ENIO RODRIGO MARCONCINI
gtalk: eniorm em gmail.com
skype: eniorm
msn: /dev/null

.: FreeBSD -:- OpenBSD -:-Slackware Linux :.
Have trouble with Windows - reboot!
Have trouble with Unix - be root!


Mais detalhes sobre a lista de discussão freebsd