Re: Too many references: cannot splice

2016-12-02 Thread Sebastian Köhler
December 2, 2016 6:58 PM, "Luis Ugalde" <forondar...@gmail.com> wrote:
> On October 13th I sent a new email with subject "Too many references:
> cannot splice" with a reference to a lklm post.

Thanks Luis for the info. I wonder why I did not find that post...

> You will find more info. and a patch there. I can confirm the patch is
> working for us.

I have applied that patch to the server and I will report back in a couple of 
days if it is also working for us.

-- Sebastian


Re: Too many references: cannot splice

2016-12-02 Thread Luis Ugalde
Sebastian,

On October 13th I sent a new email with subject "Too many references:
cannot splice" with a reference to a lklm post.

You will find more info. and a patch there. I can confirm the patch is
working for us.

Regards,
Luis Ugalde

On Fri, Dec 2, 2016 at 5:23 PM, Sebastian Köhler <s...@tyrion.de> wrote:

> Hello,
>
> multiple times per day one of our Dovecot servers has the problem that it
> suddenly refuses to accept new connections and then floods the logs with
> these messages:
>
> Dec  2 12:43:06 alfa3201 dovecot: pop3-login: Error: fd_send(pop3, 18)
> failed: Too many references: cannot splice
> Dec  2 12:43:07 alfa3201 dovecot: pop3-login: Error: fd_send(pop3, 18)
> failed: Too many references: cannot splice
> Dec  2 12:43:07 alfa3201 dovecot: imap-login: Error: fd_send(imap, 16)
> failed: Too many references: cannot splice
> Dec  2 12:43:07 alfa3201 dovecot: imap-login: Error: fd_send(imap, 16)
> failed: Too many references: cannot splice
>
>
> We have to either do a restart of the server or a doveadm kick on all
> users to make it reachable again. On my search for a solution I found a
> similar report from March[1] however without a fix.
>
>
> [1] http://dovecot.org/pipermail/dovecot/2016-March/103514.html
>
>
> dovecot -n
>
> # 2.2.24 (a82c823): /etc/dovecot/dovecot.conf
> # Pigeonhole version 0.4.13 (7b14904)
> # OS: Linux 4.4.17-040417-generic x86_64 Debian 8.6
> auth_mechanisms = plain login
> debug_log_path = /var/log/dovecot-debug.log
> default_client_limit = 8003
> default_process_limit = 2000
> default_vsz_limit = 512 M
> disable_plaintext_auth = no
> listen = *
> login_greeting = [CENSORED] - ready.
> mail_location = maildir:~/Maildir:CONTROL=/opt/dovecot/control/%n-
> control:INDEX=/opt/dovecot/index/%n
> mail_plugins = " stats quota"
> namespace inbox {
>   inbox = yes
>   location =
>   mailbox Drafts {
> auto = subscribe
> special_use = \Drafts
>   }
>   mailbox Spam {
> auto = subscribe
> special_use = \Junk
>   }
>   mailbox Sent {
> auto = subscribe
> special_use = \Sent
>   }
>   mailbox "Sent Messages" {
> special_use = \Sent
>   }
>   mailbox Trash {
> auto = subscribe
> special_use = \Trash
>   }
>   prefix = INBOX.
>   separator = .
> }
> passdb {
>   args = dovecot-%s
>   driver = pam
> }
> plugin {
>   quota = fs:%n(soft quota)
>   sieve = ~/.dovecot.sieve
>   sieve_dir = ~/sieve
>   stats_command_min_time = 1 mins
>   stats_domain_min_time = 12 hours
>   stats_ip_min_time = 12 hours
>   stats_memory_limit = 16 M
>   stats_refresh = 30 secs
>   stats_session_min_time = 15 mins
>   stats_track_cmds = no
>   stats_user_min_time = 24 hours
> }
> protocols = " imap pop3"
> service auth {
>   unix_listener /var/spool/postfix/private/auth {
> group = postfix
> mode = 0666
> user = postfix
>   }
> }
> service imap-login {
>   inet_listener imaps {
> ssl = yes
>   }
> }
> service imap-postlogin {
>   executable = script-login /etc/dovecot/post-login/expire.sh
> }
> service imap {
>   executable = imap imap-postlogin
> }
> service pop3-login {
>   inet_listener pop3 {
> port = 110
>   }
>   inet_listener pop3s {
> port = 995
> ssl = yes
>   }
> }
> service pop3-postlogin {
>   executable = script-login /etc/dovecot/post-login/expire.sh
> }
> service pop3 {
>   executable = pop3 pop3-postlogin
> }
> service quota-status {
>   executable = quota-status -p postfix
>   unix_listener /var/spool/postfix/private/quota {
> group = postfix
> mode = 0660
> user = postfix
>   }
> }
> service stats {
>   fifo_listener stats-mail {
> mode = 0666
> user = dovecot
>   }
> }
> ssl_ca =  ssl_cert =  ssl_key =  ssl_protocols = !SSLv2 !SSLv3
> userdb {
>   driver = passwd
> }
> protocol imap {
>   imap_capability = +XLIST +NAMESPACE METADATA QUOTA
>   imap_idle_notify_interval = 1 mins
>   mail_max_userip_connections = 500
>   mail_plugins = " stats quota imap_quota imap_stats"
> }
> protocol pop3 {
>   mail_max_userip_connections = 500
>   mail_plugins = " stats quota"
>   pop3_client_workarounds = outlook-no-nuls oe-ns-eoh
> }
>


Too many references: cannot splice

2016-12-02 Thread Sebastian Köhler
Hello,

multiple times per day one of our Dovecot servers has the problem that it 
suddenly refuses to accept new connections and then floods the logs with these 
messages:

Dec  2 12:43:06 alfa3201 dovecot: pop3-login: Error: fd_send(pop3, 18) failed: 
Too many references: cannot splice
Dec  2 12:43:07 alfa3201 dovecot: pop3-login: Error: fd_send(pop3, 18) failed: 
Too many references: cannot splice
Dec  2 12:43:07 alfa3201 dovecot: imap-login: Error: fd_send(imap, 16) failed: 
Too many references: cannot splice
Dec  2 12:43:07 alfa3201 dovecot: imap-login: Error: fd_send(imap, 16) failed: 
Too many references: cannot splice


We have to either do a restart of the server or a doveadm kick on all users to 
make it reachable again. On my search for a solution I found a similar report 
from March[1] however without a fix. 


[1] http://dovecot.org/pipermail/dovecot/2016-March/103514.html


dovecot -n

# 2.2.24 (a82c823): /etc/dovecot/dovecot.conf
# Pigeonhole version 0.4.13 (7b14904)
# OS: Linux 4.4.17-040417-generic x86_64 Debian 8.6
auth_mechanisms = plain login
debug_log_path = /var/log/dovecot-debug.log
default_client_limit = 8003
default_process_limit = 2000
default_vsz_limit = 512 M
disable_plaintext_auth = no
listen = *
login_greeting = [CENSORED] - ready.
mail_location = 
maildir:~/Maildir:CONTROL=/opt/dovecot/control/%n-control:INDEX=/opt/dovecot/index/%n
mail_plugins = " stats quota"
namespace inbox {
  inbox = yes
  location =
  mailbox Drafts {
auto = subscribe
special_use = \Drafts
  }
  mailbox Spam {
auto = subscribe
special_use = \Junk
  }
  mailbox Sent {
auto = subscribe
special_use = \Sent
  }
  mailbox "Sent Messages" {
special_use = \Sent
  }
  mailbox Trash {
auto = subscribe
special_use = \Trash
  }
  prefix = INBOX.
  separator = .
}
passdb {
  args = dovecot-%s
  driver = pam
}
plugin {
  quota = fs:%n(soft quota)
  sieve = ~/.dovecot.sieve
  sieve_dir = ~/sieve
  stats_command_min_time = 1 mins
  stats_domain_min_time = 12 hours
  stats_ip_min_time = 12 hours
  stats_memory_limit = 16 M
  stats_refresh = 30 secs
  stats_session_min_time = 15 mins
  stats_track_cmds = no
  stats_user_min_time = 24 hours
}
protocols = " imap pop3"
service auth {
  unix_listener /var/spool/postfix/private/auth {
group = postfix
mode = 0666
user = postfix
  }
}
service imap-login {
  inet_listener imaps {
ssl = yes
  }
}
service imap-postlogin {
  executable = script-login /etc/dovecot/post-login/expire.sh
}
service imap {
  executable = imap imap-postlogin
}
service pop3-login {
  inet_listener pop3 {
port = 110
  }
  inet_listener pop3s {
port = 995
ssl = yes
  }
}
service pop3-postlogin {
  executable = script-login /etc/dovecot/post-login/expire.sh
}
service pop3 {
  executable = pop3 pop3-postlogin
}
service quota-status {
  executable = quota-status -p postfix
  unix_listener /var/spool/postfix/private/quota {
group = postfix
mode = 0660
user = postfix
  }
}
service stats {
  fifo_listener stats-mail {
mode = 0666
user = dovecot
  }
}
ssl_ca = 

Re: Too many references: cannot splice

2016-11-08 Thread Luis Ugalde
Hi,

I've been running the patched version for the last few days, and I can
confirm that I have not seen any errors. I still have to check it with 4.x
kernels, but it's working as expected with the current Debian Stable one.


Regards,
Luis ugalde.

On Wed, Oct 26, 2016 at 12:59 PM, Timo Sirainen <t...@iki.fi> wrote:

> On 26 Oct 2016, at 11:14, Luis Ugalde <forondar...@gmail.com> wrote:
> >
> > Hi,
> >
> > Could you please have a look at  https://lkml.org/lkml/2016/2/2/538 and
> see
> > if this makes any sense to you? I've been checking kernel changes
> > between linux_3.16.7 and linux_3.16.36, and this has popped out. Could
> this
> > be the reason for the "too many references" errors?
>
> Does the attached patch help?
>
>
>
> >
> > Regards,
> >
> > Luis Ugalde.
> >
> > On Thu, Oct 13, 2016 at 3:47 PM, Luis Ugalde <forondar...@gmail.com>
> wrote:
> >
> >> Hi,
> >>
> >>
> >> A while ago I sent an email regarding these "*ETOOMANYREFS* Too many
> >> references: cannot splice." that we've seen since Debian updated the
> Jessie
> >> kernel to
> >>
> >> 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17) x86_64
> >>
> >> while older kernels, like  3.16.0-4-amd64 #1 SMP Debian
> >> 3.16.7-ckt11-1+deb8u6 (2015-11-09) x86_64 showed no errors at all.
> >>
> >> I was wondering if no one else is getting these errors, or if you know
> any
> >> workarounds that might probe useful, apart from downgrading the kernel.
> >>
> >>
> >> I would say that the infrastructure we're running is quite standard,
> with
> >> directors balancing users to NFS backed dovecot servers.
> >>
> >>
> >> Best regards,
> >>
> >> Luis Ugalde.
> >>
> >>
> >>
> >>
>
>
>


Re: Too many references: cannot splice

2016-10-26 Thread Timo Sirainen
On 26 Oct 2016, at 11:14, Luis Ugalde <forondar...@gmail.com> wrote:
> 
> Hi,
> 
> Could you please have a look at  https://lkml.org/lkml/2016/2/2/538 and see
> if this makes any sense to you? I've been checking kernel changes
> between linux_3.16.7 and linux_3.16.36, and this has popped out. Could this
> be the reason for the "too many references" errors?

Does the attached patch help?



diff
Description: Binary data

> 
> Regards,
> 
> Luis Ugalde.
> 
> On Thu, Oct 13, 2016 at 3:47 PM, Luis Ugalde <forondar...@gmail.com> wrote:
> 
>> Hi,
>> 
>> 
>> A while ago I sent an email regarding these "*ETOOMANYREFS* Too many
>> references: cannot splice." that we've seen since Debian updated the Jessie
>> kernel to
>> 
>> 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17) x86_64
>> 
>> while older kernels, like  3.16.0-4-amd64 #1 SMP Debian
>> 3.16.7-ckt11-1+deb8u6 (2015-11-09) x86_64 showed no errors at all.
>> 
>> I was wondering if no one else is getting these errors, or if you know any
>> workarounds that might probe useful, apart from downgrading the kernel.
>> 
>> 
>> I would say that the infrastructure we're running is quite standard, with
>> directors balancing users to NFS backed dovecot servers.
>> 
>> 
>> Best regards,
>> 
>> Luis Ugalde.
>> 
>> 
>> 
>> 



Re: Too many references: cannot splice

2016-10-26 Thread Luis Ugalde
Hi,

Could you please have a look at  https://lkml.org/lkml/2016/2/2/538 and see
if this makes any sense to you? I've been checking kernel changes
between linux_3.16.7 and linux_3.16.36, and this has popped out. Could this
be the reason for the "too many references" errors?

Regards,

Luis Ugalde.

On Thu, Oct 13, 2016 at 3:47 PM, Luis Ugalde <forondar...@gmail.com> wrote:

> Hi,
>
>
> A while ago I sent an email regarding these "*ETOOMANYREFS* Too many
> references: cannot splice." that we've seen since Debian updated the Jessie
> kernel to
>
> 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17) x86_64
>
> while older kernels, like  3.16.0-4-amd64 #1 SMP Debian
> 3.16.7-ckt11-1+deb8u6 (2015-11-09) x86_64 showed no errors at all.
>
> I was wondering if no one else is getting these errors, or if you know any
> workarounds that might probe useful, apart from downgrading the kernel.
>
>
> I would say that the infrastructure we're running is quite standard, with
> directors balancing users to NFS backed dovecot servers.
>
>
> Best regards,
>
> Luis Ugalde.
>
>
>
>


Too many references: cannot splice

2016-10-13 Thread Luis Ugalde
Hi,


A while ago I sent an email regarding these "*ETOOMANYREFS* Too many
references: cannot splice." that we've seen since Debian updated the Jessie
kernel to

3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt20-1+deb8u3 (2016-01-17) x86_64

while older kernels, like  3.16.0-4-amd64 #1 SMP Debian
3.16.7-ckt11-1+deb8u6 (2015-11-09) x86_64 showed no errors at all.

I was wondering if no one else is getting these errors, or if you know any
workarounds that might probe useful, apart from downgrading the kernel.


I would say that the infrastructure we're running is quite standard, with
directors balancing users to NFS backed dovecot servers.


Best regards,

Luis Ugalde.