[Dovecot] very slow pop3 downloads
Hi all,
I have a dovecot installation which very suddenly decided to slow to a crawl yesterday after being totally fine for months.
IMAP is fine, but POP3 is very painful to use, and frequently times out.
here's my config: # 1.0.7: /etc/dovecot.conf log_path: /var/log/dovecot_errors info_log_path: /var/log/dovecot_informational ssl_ca_file: /etc/postfix/ssl/cacert.pem ssl_cert_file: /etc/postfix/ssl/postfix.webworks.ie.crt ssl_key_file: /etc/postfix/ssl/postfix.webworks.ie.key login_dir: /var/run/dovecot/login login_executable(default): /usr/libexec/dovecot/imap-login login_executable(imap): /usr/libexec/dovecot/imap-login login_executable(pop3): /usr/libexec/dovecot/pop3-login login_process_per_connection: no login_processes_count: 6 first_valid_uid: 89 mail_location: maildir:/home/vmail/%d/%n mail_executable(default): /usr/libexec/dovecot/imap mail_executable(imap): /usr/libexec/dovecot/imap mail_executable(pop3): /usr/libexec/dovecot/pop3 mail_plugin_dir(default): /usr/lib/dovecot/imap mail_plugin_dir(imap): /usr/lib/dovecot/imap mail_plugin_dir(pop3): /usr/lib/dovecot/pop3 pop3_client_workarounds(default): pop3_client_workarounds(imap): pop3_client_workarounds(pop3): outlook-no-nuls oe-ns-eoh auth default: passdb: driver: pam passdb: driver: sql args: /etc/dovecot-mysql.conf userdb: driver: sql args: /etc/dovecot-mysql.conf
I tried an strace as suggested by some other people. didn't know exactly what to attach it to, so did a "ps aux" (Centos 5.2, btw) and chose the first "pop3" process:
[root@postfix cur]# strace -p 1697 Process 1697 attached - interrupt to quit gettimeofday({1238146840, 724484}, {0, 0}) = 0 gettimeofday({1238146840, 724542}, NULL) = 0 poll([{fd=4, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=2, events=POLLERR|POLLHUP|POLLNVAL}, {fd=0, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=1, events=POLLOUT|POLLERR|POLLHUP|POLLNVAL}], 4, 3123) = 0 gettimeofday({1238146843, 847540}, {0, 0}) = 0 gettimeofday({1238146843, 847590}, NULL) = 0 poll([{fd=4, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=2, events=POLLERR|POLLHUP|POLLNVAL}, {fd=0, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=1, events=POLLOUT|POLLERR|POLLHUP|POLLNVAL, revents=POLLOUT}], 4, 10000) = 1 gettimeofday({1238146847, 807077}, {0, 0}) = 0 write(1, "6DEDK02J5VXFQtloRQ/JPgdcCrSmSRXj"..., 2171) = 2171 setsockopt(1, SOL_TCP, TCP_CORK, [1], 4) = 0 pread64(9, "7P8JZloV9g/oYWdnm\nnZ9f+xzkO3fVXr"..., 4096, 102400) = 4096 write(1, "\r\nLhQoxZy+rq1TZXl2ZtrKrZH15dE2PK"..., 4134) = 3669 write(1, "uv69IdP6FDbO0lztPh2sA6Z/wBqIdP9Y"..., 465) = -1 EAGAIN (Resource temporarily unavailable) setsockopt(1, SOL_TCP, TCP_CORK, [0], 4) = 0 gettimeofday({1238146847, 807691}, NULL) = 0 poll([{fd=4, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=2, events=POLLERR|POLLHUP|POLLNVAL}, {fd=0, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=1, events=POLLOUT|POLLERR|POLLHUP|POLLNVAL}], 4, 6040) = 0 gettimeofday({1238146853, 847046}, {0, 0}) = 0 gettimeofday({1238146853, 847095}, NULL) = 0 poll([{fd=4, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=2, events=POLLERR|POLLHUP|POLLNVAL}, {fd=0, events=POLLIN|POLLPRI|POLLERR|POLLHUP|POLLNVAL}, {fd=1, events=POLLOUT|POLLERR|POLLHUP|POLLNVAL, revents=POLLOUT}], 4, 10000) = 1 gettimeofday({1238146856, 546780}, {0, 0}) = 0 write(1, "uv69IdP6FDbO0lztPh2sA6Z/wBqIdP9Y"..., 465) = 465 setsockopt(1, SOL_TCP, TCP_CORK, [1], 4) = 0 pread64(9, "24\nekSIYgeKr5kpQwrk4iWMqC2+/uvwf"..., 4096, 106496) = 4096 write(1, "\r\nIHaoJDSPFqr4jw9SP8S2yLsWOt2RVx"..., 4134) = 4134 pread64(9, "DUJvKTRdVNxs3JTdRQYqD61EYbg22DLb"..., 4096, 110592) = 4096 write(1, "\r\nbDlx3+0LszvjOrD5+04T8PF39frFpQ"..., 4134) = 1241 write(1, "9Zc/wAr\r\nF1bI22U8dNTIKBc423KFJSO"..., 2893) = -1 EAGAIN (Resource temporarily unavailable) setsockopt(1, SOL_TCP, TCP_CORK, [0], 4) = 0 gettimeofday({1238146856, 547406}, NULL) = 0 poll( <unfinished ...> Process 1697 detached
looking forward to you all saving my sanity! thanks, kae
Kae Verens wrote:
Hi all,
I have a dovecot installation which very suddenly decided to slow to a crawl yesterday after being totally fine for months.
sorry for the immediate reply to my own request for help.
I just found out that the problem may not be on the machine - our upstream supplier is having the same problem on one of his mail servers, which is in the same datacentre.
this may not be a problem with dovecot - sorry for the waste of time!
kae
participants (1)
-
Kae Verens