Browser freezes on grouped messages

[josebarros]

Hi all,
Good morning.

I'm facing an odd issue with my browser when accessing grouped messages on webmail.

When the thread is bigger than 50 messages, all or most of them unread, my browsers freezes for a while.

My assumption is that this happens while marking all messages as read, although the "round" number 50 is a turn over fort his to happen.

Is there anything we may configure on Zimbra that may avoid this?

Looking forward to hearing from you.

Thank you in advance.

Best regards,
JB

[pup_seba]

Hi,

The info you see in the browser navigation pane, it all comes from the database. Zimbra only accesses the message itself once you clic on one message. For the case of conversations, I imagine that you are forcing zimbra to read from disc from "50 messages" at the time.

Meaning 2 things:
- either you refer to slowliness at the db level, which is due to probably a bad configured innodb_buffer size (compare "du -sh /opt/zimbra/db/data" with the size assigned to the innodb_buffer_size in /opt/zimbra/conf/my.cnf)
- you don't have a fast enough disk backing your operations. It is very common to see people using RAID5 or even SATA disks or 1gb nfs mounted ones to hold their "store".

Regards,

[josebarros]

Hi pup_seba, thank you for your response.
Just for curiosity, just noticed we have joined this forum on the same day.

Hi,

The info you see in the browser navigation pane, it all comes from the database. Zimbra only accesses the message itself once you clic on one message. For the case of conversations, I imagine that you are forcing zimbra to read from disc from "50 messages" at the time.

I've never changed any rule related to the number of messages to read.
Imagine one scenario where, during the night, I received multiple log emails, with the same subject from the same sender.
All message will be grouped. If they are 60, for example, when I click on the group conversation, the browser hangs and after a while unfreezes with all messages marked as read. On slower computers, the freeze is a bit too long, sometimes we have to close the browser...

Meaning 2 things:
- either you refer to slowliness at the db level, which is due to probably a bad configured innodb_buffer size (compare "du -sh /opt/zimbra/db/data" with the size assigned to the innodb_buffer_size in /opt/zimbra/conf/my.cnf)

The du command shows the following data:

Code: Select all

du -sh /opt/zimbra/db/data
6.0G	/opt/zimbra/db/data

But I cannot find any innodb_buffer_size on the /opt/zimbra/conf/my.cnf file.

Here is my.cnf file contents:

Code: Select all

[mysqld]

bind-address = 127.0.0.1






basedir        = /opt/zimbra/common
datadir        = /opt/zimbra/db/data
socket         = /opt/zimbra/data/tmp/mysql/mysql.sock
pid-file       = /opt/zimbra/log/mysql.pid
port           = 7306
user           = zimbra
tmpdir         = /opt/zimbra/data/tmp

external-locking
slow_query_log = 1
slow_query_log_file = /opt/zimbra/log/myslow.log

general_log_file = /opt/zimbra/log/mysql-mailboxd.log

long_query_time  = 1
log_queries_not_using_indexes

thread_cache_size = 110
max_connections   = 110

# We do a lot of writes, query cache turns out to be not useful.
query_cache_type = 0

sort_buffer_size = 1048576
read_buffer_size = 1048576

# (Num mailbox groups * Num tables in each group) + padding
table_open_cache = 1200

innodb_data_file_path          = ibdata1:10M:autoextend
innodb_buffer_pool_size        = 3094732800
innodb_log_file_size           = 524288000
innodb_log_buffer_size         = 8388608
innodb_file_per_table

# Value is: 200 + max_connections + 2 * table_open_cache
innodb_open_files              = 2710

innodb_max_dirty_pages_pct     = 30
innodb_flush_method            = O_DIRECT
innodb_flush_log_at_trx_commit = 0
max_allowed_packet             = 16777216

[mysqld_safe]

log-error      = /opt/zimbra/log/mysqld.log
pid-file     = /opt/zimbra/log/mysql.pid

Should I add it?
With which value?

- you don't have a fast enough disk backing your operations. It is very common to see people using RAID5 or even SATA disks or 1gb nfs mounted ones to hold their "store".

This should not be the issue. We are running on Intel Enterprise NvME disk files in Raid1.
The mailserver also has plenty of RAM (16GB) to support ~40 mail accounts.

Looking forward for your feedback.

Best regards,
JB