Skip to main content

Hadoop `sysctl.conf` parameters

Hadoop sysctl.conf parameters.
File System
  1. fs.file-max Increase size of file handles and inode cache
 [ahmed@server ~]# echo 'fs.file-max = 943718' >> /etc/sysctl.conf
Swappiness : Do less swapping
  1. vm.dirty_ratio setting virtual memory ratio.
  2. vm.swappiness How often swap should be used. 0 is least, 60 default.
 [ahmed@server ~]# echo 'vm.dirty_ratio=10' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'vm.swappiness=0' >> /etc/sysctl.conf
Connection Settings
  1. net.core.netdev_max_backlog Increase number of incoming connections backlog.
  2. net.core.somaxconn Increase number of incoming connections.
 [ahmed@server ~]# echo 'net.core.netdev_max_backlog = 4000' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.core.somaxconn = 4000' >> /etc/sysctl.conf
TCP settings
  1. net.ipv4.tcp_sack Disable select acknowledgments
  2. net.ipv4.tcp_dsack Allows TCP to send “duplicate” SACKs.
  3. net.ipv4.tcp_keepalive_time How often TCP sends out keepalive messages when keepalive is enabled. Default: 2hours.
  4. net.ipv4.tcp_keepalive_probes How many keepalive probes TCP sends out, until it decides that the connection is broken. Default value: 9.
  5. net.ipv4.tcp_keepalive_intvl How frequently the probes are send out. Multiplied by tcp_keepalive_probes it is time to kill not responding connection, after probes started. Default value: 75sec i.e. connection will be aborted after ~11 minutes of retries.
  6. net.ipv4.tcp_fin_timeout Time to hold socket in state FIN-WAIT-2, if it was closed by our side. Peer can be broken and never close its side, or even died unexpectedly. Default value is 60sec. Usual value used in 2.2 was 180 seconds, you may restore it, but remember that if your machine is even underloaded WEB server, you risk to overflow memory with kilotons of dead sockets, FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1, because they eat maximum 1.5K of memory, but they tend to live longer. Cf.tcp_max_orphans.
  7. net.ipv4.tcp_rmem The three values setting the minimum, initial, and maximum size of the Memory Receive Buffer per connection. They define the actual memory usage, not just TCP window size.
  8. net.ipv4.tcp_wmem The same as tcp_rmem, but just for Memory Send Buffer per connection.
  9. net.ipv4.tcp_retries2 This value influences the timeout of an alive TCP connection, when RTO retransmissions remain unacknowledged. Given a value of N, a hypothetical TCP connection following exponential backoff with an initial RTO of TCP_RTO_MIN would retransmit N times before killing the connection at the (N+1)th RTO. The default value of 15 yields a hypothetical timeout of 924.6 seconds and is a lower bound for the effective timeout. TCP will effectively time out at the first RTO which exceeds the hypothetical timeout. RFC 1122 recommends at least 100 seconds for the timeout, which corresponds to a value of at least 8.
  10. net.ipv4.tcp_synack_retries Number of times SYNACKs for a passive TCP connection attempt will be retransmitted. Should not be higher than 255. Default value is 5, which corresponds to ~180seconds.
 [ahmed@server ~]# echo 'net.ipv4.tcp_sack = 0' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_dsack = 0' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_keepalive_time = 600' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_keepalive_probes = 5' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_keepalive_intvl = 15' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_fin_timeout = 30' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_rmem = 32768 436600 4194304' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_wmem = 32768 436600 4194304' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_retries2 = 10' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv4.tcp_synack_retries = 3' >> /etc/sysctl.conf
Disable IPv6 Defaults. We dont use these anyway.
 [ahmed@server ~]# echo 'net.ipv6.conf.all.disable_ipv6 = 1' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv6.conf.default.disable_ipv6 = 1' >> /etc/sysctl.conf
 [ahmed@server ~]# echo 'net.ipv6.conf.lo.disable_ipv6 = 1' >> /etc/sysctl.conf
Execute below command to make it permanent.
 [ahmed@server ~]# sysctl -p
Next update limits.
 [ahmed@server ~]# echo '* - nofile 65536' >>/etc/security/limits.conf
 [ahmed@server ~]# echo '* - nproc 65536' >>/etc/security/limits.conf
More Details on IPv4
 http://www.cyberciti.biz/files/linux-kernel/Documentation/networking/ip-sysctl.txt
 http://www.cyberciti.biz/files/linux-kernel/Documentation/sysctl/
​
Labels:

Comments

Post a Comment

Popular posts from this blog

Cloudera Manager - Duplicate entry 'zookeeper' for key 'NAME'.

We had recently built a cluster using cloudera API’s and had all the services running on it with Kerberos enabled. Next we had a requirement to add another kafka cluster to our already exsisting cluster in cloudera manager. Since it is a quick task to get the zookeeper and kafka up and running. We decided to get this done using the cloudera manager instead of the API’s. But we faced the Duplicate entry 'zookeeper' for key 'NAME' issue as described in the bug below. https://issues.cloudera.org/browse/DISTRO-790 I have set up two clusters that share a Cloudera Manger. The first I set up with the API and created the services with capital letter names, e.g., ZOOKEEPER, HDFS, HIVE. Now, I add the second cluster using the Wizard. Add Cluster->Select Hosts->Distribute Parcels->Select base HDFS Cluster install On the next page i get SQL errros telling that the services i want to add already exist. I suspect that the check for existing service names does n
2 comments
Read more

Zabbix History Table Clean Up

Zabbix history table gets really big, and if you are in a situation where you want to clean it up. Then we can do so, using the below steps. Stop zabbix server. Take table backup - just in case. Create a temporary table. Update the temporary table with data required, upto a specific date using epoch . Move old table to a different table name. Move updated (new temporary) table to original table which needs to be cleaned-up. Drop the old table. (Optional) Restart Zabbix Since this is not offical procedure, but it has worked for me so use it at your own risk. Here is another post which will help is reducing the size of history tables - http://zabbixzone.com/zabbix/history-and-trends/ Zabbix Version : Zabbix v2.4 Make sure MySql 5.1 is set with InnoDB as innodb_file_per_table=ON Step 1 Stop the Zabbix server sudo service zabbix-server stop Script. echo "------------------------------------------" echo " 1. Stopping Zabbix Server &quo
1 comment
Read more

Access Filter in SSSD `ldap_access_filter` [SSSD Access denied / Permission denied ]

Access Filter Setup with SSSD ldap_access_filter (string) If using access_provider = ldap , this option is mandatory. It specifies an LDAP search filter criteria that must be met for the user to be granted access on this host. If access_provider = ldap and this option is not set, it will result in all users being denied access. Use access_provider = allow to change this default behaviour. Example: access_provider = ldap ldap_access_filter = memberOf=cn=allowed_user_groups,ou=Groups,dc=example,dc=com Prerequisites yum install sssd Single LDAP Group Under domain/default in /etc/sssd/sssd.conf add: access_provider = ldap ldap_access_filter = memberOf=cn=Group Name,ou=Groups,dc=example,dc=com Multiple LDAP Groups Under domain/default in /etc/sssd/sssd.conf add: access_provider = ldap ldap_access_filter = (|(memberOf=cn=System Adminstrators,ou=Groups,dc=example,dc=com)(memberOf=cn=Database Users,ou=Groups,dc=example,dc=com)) ldap_access_filter accepts standa
Post a Comment
Read more