Selasa, 18 September 2018

Recomendasi server owncloud

Deployment Recommendations

What is the best way to install and maintain ownCloud? The answer to that is, as always: “it depends”.
This is because every ownCloud customer has their own particular needs and IT infrastructure. However, both ownCloud and the LAMP stack are highly-configurable. Given that, in this document we present a set of general recommendations, followed by three typical scenarios, and finish up with making best-practice recommendations for both software and hardware.
The recommendations presented here are based on a standard ownCloud installation, one without any particular appsthemes, or code changes. But, server load is dependent upon the number of clientsfiles, and user activity, as well as other usage patterns. Given that, these recommendations are only a rule of thumb based on our experience, as well as that of one of our customers.

General Recommendations

  • Operating system: Linux.
  • Web server: Apache 2.4.
  • Database: MySQL/MariaDB with InnoDB storage engine (MyISAM is not supported, see: MySQL / MariaDB storage engine)
  • PHP 5.6+.
  • Consider setting up a scale-out deployment, or using Federated Cloud Sharing to keep individual ownCloud instances to a manageable size.
Whatever the size of your organization, always keep one thing in mind: The amount of data stored in ownCloud will only grow. So plan ahead.

ownCloud Administrators Must Have Command Line or Cron Access

We only recommend using hosts that provide command-line or Cron access (ideally both) to ownCloud administrators, for three key reasons:
  1. Without command-line access, OCC commands, required for administrative tasks such as repairs and upgrades, are not available.
  2. Without crontab access, you cannot run background jobs reliably. ajax/cron.php is available, but it is not reliable enough, because it only runs when people are using the web UI. Additionally, ownCloud relies heavily on background jobs especially for long-running operations, which will likely cause PHP timeouts.
  3. PHP timeout values are often low. Having low timeout settings can break long-running operations, such as moving a huge folder.

Scenario 1: Small Workgroups and Departments

This recommendation applies if you meet the following criteria:
OptionValue
Number of usersUp to 150 users
Storage size100 GB to 10TB
High availability levelZero-downtime backups via Btrfs snapshots, component failure leads to interruption of service. Alternate backup scheme on other filesystems: nightly backups — with service interruption.

Scenario 2: Mid-Sized Enterprises

These recommendations apply if you meet the following criteria:
OptionValue
Number of users150 to 1,000 users.
Storage sizeUp to 200TB.
High availability levelEvery component is fully redundant and can fail without service interruption. Backups without service interruption

Recommended System Requirements

  • 2 to 4 application servers.
  • A cluster of two database servers.
  • Storage on an NFS server.
  • Authentication via an existing LDAP or Active Directory server.
Network diagram for a mid-sized enterprise.

Components

  • 2 to 4 application servers with four sockets and 32GB RAM.
  • 2 DB servers with four sockets and 64GB RAM.
  • HAproxy load balancer with two sockets and 16GB RAM.
  • NFS storage server as needed.

Operating System

Enterprise grade Linux distribution with full support from an operating system vendor. We recommend both RedHat Enterprise Linux and SUSE Linux Enterprise Server 12.

SSL Configuration

The SSL termination is done in the HAProxy load balancer. A standard SSL certificate is needed, installed according to the HAProxy documentation.

Load Balancer

HAProxy running on a dedicated server in front of the application servers. Sticky session needs to be used because of local session management on the application servers.

Database

MySQL/MariaDB Galera cluster with master-master replication. InnoDB storage engine, MyISAM is not supported, see: MySQL / MariaDB storage engine.

Backup

Minimum daily backup without downtime. All MySQL/MariaDB statements should be replicated to a backup MySQL/MariaDB slave instance.
  • Create a snapshot on the NFS storage server.
  • At the same time stop the MySQL replication.
  • Create a MySQL dump of the backup slave.
  • Push the NFS snapshot to the backup.
  • Push the MySQL dump to the backup.
  • Delete the NFS snapshot.
  • Restart MySQL replication.

Authentication

User authentication via one or several LDAP or Active Directory servers. See User Authentication with LDAP for information on configuring ownCloud to use LDAP and AD.

Session Management

Session management on the application server. PHP sessions are stored in a temporary filesystem, mounted at the operating system-specific session storage location. You can find out where that is by running grep -R 'session.save_path' /etc/php5 and then add it to the /etc/fstab file, for example:
echo "tmpfs /var/lib/php5/pool-www tmpfs defaults,noatime,mode=1777 0 0" >> /etc/fstab

Memory Caching

A memory cache speeds up server performance, and ownCloud supports four memory cache types. Refer to Configuring Memory Caching for information on selecting and configuring a memory cache.

Storage

Use an off-the-shelf NFS solution, such as IBM Elastic Storage or RedHat Ceph.

ownCloud Edition

Enterprise Edition. See ownCloud Server or Enterprise Edition for comparisons of the ownCloud editions.

Scenario 3: Large Enterprises and Service Providers

OptionValue
Number of users5,000 to >100,000 users.
Storage sizeUp to 1 petabyte.
High availability levelEvery component is fully redundant and can fail without service interruption. Backups without service interruption.

Recommended System Requirements

  • 4 to 20 application/Web servers.
  • A cluster of two or more database servers.
  • Storage is an NFS server or an object store that is S3 compatible.
  • Cloud federation for a distributed setup over several data centers.
  • Authentication via an existing LDAP or Active Directory server, or SAML.
Network diagram for large enterprise.

Components

  • 4 to 20 application servers with four sockets and 64GB RAM.
  • 4 DB servers with four sockets and 128GB RAM.
  • 2 Hardware load balancer, for example, BIG IP from F5.
  • NFS storage server as needed.

Operating system

RHEL 7 with latest service packs.

SSL Configuration

The SSL termination is done in the load balancer. A standard SSL certificate is needed, installed according to the load balancer documentation.

Load Balancer

A redundant hardware load-balancer with heartbeat, for example, F5 Big-IP. This runs two load balancers in front of the application servers.

Database

MySQL/MariaDB Galera Cluster with 4x master-master replication. InnoDB storage engine, MyISAM is not supported, see: MySQL / MariaDB storage engine.

Backup

Minimum daily backup without downtime. All MySQL/MariaDB statements should be replicated to a backup MySQL/MariaDB slave instance. To do this, follow these steps:
  1. Create a snapshot on the NFS storage server.
  2. At the same time stop the MySQL replication.
  3. Create a MySQL dump of the backup slave.
  4. Push the NFS snapshot to the backup.
  5. Push the MySQL dump to the backup.
  6. Delete the NFS snapshot.
  7. Restart MySQL replication.

Authentication

User authentication via one or several LDAP or Active Directory servers, or SAML/Shibboleth. See User Authentication with LDAP and Shibboleth Integration.

LDAP

Read-only slaves should be deployed on every application server for optimal scalability.

Session Management

Redis should be used for the session management storage.

Caching

Redis for distributed in-memory caching, see Configuring Memory Caching.

Storage

An off-the-shelf NFS solution should be used. Some examples are IBM Elastic Storage or RedHat Ceph. Optionally, an S3 compatible object store can also be used.

ownCloud Edition

Enterprise Edition. See ownCloud Server or Enterprise Edition for comparisons of the ownCloud editions.

Redis Configuration

Redis in a master-slave configuration is a hot failover setup, and is usually sufficient. A slave can be omitted if high availability is provided via other means. And when it is, in the event of a failure, restarting Redis typically occurs quickly enough. Regarding Redis cluster, we don’t, usually, recommend it, as it requires a greater level of both maintenance and management in the case of failure. A single Redis server, however, just needs to be rebooted, in the event of failure.

Known Issues

Deadlocks When Using MariaDB Galera Cluster

If you’re using MariaDB Galera Cluster with your ownCloud installation, you may encounter deadlocks when you attempt to sync a large number of files. You may also encounter database errors, such as this one:
SQLSTATE[40001]: Serialization failure: 1213 Deadlock found when trying to get lock; try restarting transaction
The issue, identified by Michael Roth, is caused when MariaDB Galera cluster sends write requests to all servers in the cluster; here is a detailed explanation. The solution is to send all write requests to a single server, instead of all of them.

Set wsrep_sync_wait to 1 on all Galera Cluster nodes

What the parameter does

When enabled, the node triggers causality checks in response to certain types of queries. During the check, the node blocks new queries while the database server catches up with all updates made in the cluster to the point where the check begun. Once it reaches this point, the node executes the original query.

Why enable it

A Galera Cluster write operation is sent to the master while reads are retrieved from the slaves. Since Galera Cluster replication is, by default, not strictly synchronous it could happen that items are requested before the replication has actually taken place.
This setting is disabled by default.
See the Galera Cluster WSREP documentation for more details.

Tidak ada komentar:

Posting Komentar