Owncloud 10.0.0 is slow


#1

I passed

FROM

owncloud 9.1
php 7.0
mariadb 10.1
centos 7

TO

owncloud 10.0.0
php 7.1.5
mariadb 10.1
centos 7
redis and phpredis

CONFIG.php

....
'filelocking.enabled' => true,
'memcache.locking' => '\OC\Memcache\Redis',
'memcache.local' => '\OC\Memcache\Redis',
'redis' =>
array (
'host' => 'localhost',
'port' => 6379,
'timeout' => 0.0,
'password' => '',
),

and the response by timer is owc 10.0.0 is more slow than owx 9.1 ... it is very ugly for my client, mostly the hardware resources are same. Do exist a solution? Can I dwongrade to 9.1 without lost datas from DB or files?
thanks

All controls have passed


#2

not reply?
strange that nobody have my self issue... from 9.1.4 to 10.0.0 the owncloud for equality resources is more slow in sync (client from and to server, the icon cloud always is blue), but a lot slow than 9.1.4.... why? and mostly how do I resolve?

I notice that the check before sync is slow, after this time the owncloud becomes normal...

thanks


#3

Make sure you're running the latest version of software. What about the memory on the host? Is it writing to swap at all?


#4

owncloud installed yesterday: 10.0.2.1 but nothing it's changed. Redis's installed correctly.
the client have memory sufficient.
the strange is that owncloud 9.1.4 was more speed in sync step than 10, so I think the software owncloud 10 have some "bottleneck".
I don't understand the swap... Can you explain me?
thanks
look the link:


#5

The memory on the host (i.e. the server) is more what you should be concerned with. If your database is writing to virtual memory (swap), then that will slow things down. If you reboot the server, does the problem go away?


#6

no, I reboot more time, but the slow rest.... is it possible so many select per seconds?
how do I configure my.cnf in order to improve the performance?
now I have:

table_open_cache=1000000
innodb_io_capacity = 2000
innodb_read_io_threads = 64
innodb_thread_concurrency = 0
innodb_write_io_threads = 64
innodb_buffer_pool_size=14G
innodb_log_buffer_size = 256M
query_cache_size = 0

thanks


#7

The output from something like the 'free' command would be more useful:

free -h


#8

total used free shared buff/cache available
Mem: 30G 3.9G 400M 112M 26G 25G
Swap: 3.5G 0B 3.5G

is it useful? Can you help me?
thanks


#9

is it possible only to me the owncloud 10 runs slow?


#10

Can you specify what exactly is slow? The webinterface? Up- and download? And to what degree, can you give numbers?


#11

the slowdown is the "check before sync, checking of the changes" (after this operation anything is normal, download/upload file):

and if I look the performance server I have this:

many select per seconds and many traffic... (redis is installed correctly).
I don't understand why the same hardware/software resouces from 9.1 to 10.0 have slowed down owncloud.
thanks


#12

@PVince81 any changes in the code that could be responsible for that? Bug?


#13

from me absolutely no one, from owncloud would be help for me.

What I see is the DB work a lot, more than 10000 select per seconds are a lot... the cpu work only for DB.


#14

It would be worth configuring the mariadb server to produce a query log and checkout what are these 10000 queries; likely that they will all be quite similar and point us to where the problem lies here.


#15

You could try running a user file scan for the user that is having the issues to ensure the cache is up to date. https://doc.owncloud.org/server/9.0/admin_manual/configuration_server/occ_command.html#file-operations-label


#16

I setted my.cnf and I haven't slow query and the query are:

query.log

But today owncloud work less because today isn't working day.
Tomorrow owncloud work more than today, so I will obtain more query info and I'll send you
thanks


#17

I done... I check it


#18

Not really. OC 10.0 should actually be faster thanks to many improvements.


#19

you consider my owc 10 from upgrade owc 9.1.4, maybe have this step created slow "checking update"?
if I install a new CPU from 2.5Ghz to 3Ghz, do I speed the process?
thanks


#20

I have just upgraded from 9.1 to 10 and cant believe how much slower it is. The 'hardware;' hasn't changed. We are using PostgreSQL rather than mysql and I can't see anything obvious in terms of db issues. I moved /opt/otrs/var/tmp to a RAM disk and that has helped a little but it is pretty painful to use at present.