5,181 reputation
1133
bio website alexclear.livejournal.com
location Mother Russia
age 36
visits member for 3 years, 10 months
seen yesterday

An experienced software developer with a bit of sysadmin background. Can't dance.


Aug
23
answered How can I scroll back the output in FreeBSD's console without Scroll Lock?
Aug
23
answered Is it possible to configure CPanel to run only eith Nginx
Aug
22
answered Setup two NICs, one public and one private
Aug
22
answered With HTTPS-Enforcer chrome extension, a normal ping is answered by different site
Aug
22
comment With HTTPS-Enforcer chrome extension, a normal ping is answered by different site
This is weird, could you please post the content of your /etc/resolv.conf?
Aug
22
answered Does apache apply directives sequentially?
Aug
22
answered Analysing Disk IO Traffic
Aug
22
comment Nginx + Unicorn
There should be a staging.log file somewhere (I guess in /data/appname/staging/current/log/staging.log), it looks like your application fails before producing any output. It's also good to check unicorn.stderr.log too.
Aug
22
comment Nginx + Unicorn
What version of nginx are you using?
Aug
22
comment Nginx + Unicorn
Is there anything relevant in /data/appname/staging/current/log/unicorn-staging-error.log?
Aug
21
answered PHP occasionally fails to connect to mysql server
Aug
21
answered Unable to boot Rackspace Cloud image on KVM
Aug
21
answered Replicate GIT repository over multiple geolocations
Aug
21
comment MySQL: 5000 core users, 2GB Table Heavy Write / Read: How to Prevent Crashing
When InnoDB crashes it just replays transaction log records on restart, this is much faster than rechecking all the table. A number of transaction log records to replay after crash depends on write load only, not on the total DB size.
Aug
21
comment MySQL: 5000 core users, 2GB Table Heavy Write / Read: How to Prevent Crashing
I never use MyISAM in production especially for large setups because it's non-transactional and requires the recheck and repair procedure almost every time it crashes so I am not a big expert here. But there should be something relevant in the MySQL error log or maybe even in /var/log/syslog, many distributions log MySQL events right there.
Aug
21
comment Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) on subsequent connection through terminal
Well if the server is running and if you can connect to it after editing the [client] section of /etc/mysql/my.cnf you should probably leave everything as is.
Aug
21
comment MySQL: 5000 core users, 2GB Table Heavy Write / Read: How to Prevent Crashing
Is there anything useful in MySQL error log and slow log? Basically I would recommend you to convert everything to InnoDB but this implies a certain downtime which can be quite high (well, it should be comparable to the full check time). But if you use native MySQL full text search the conversion won't be possible because InnoDB does not support FTS.
Aug
21
comment MySQL: 5000 core users, 2GB Table Heavy Write / Read: How to Prevent Crashing
Please post your my.cnf somewhere. What storage engine do you use, MyISAM or InnoDB?
Aug
21
comment Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) on subsequent connection through terminal
I recommend you to perform a full database dump and reinstall MySQL from packages.
Aug
21
comment Can't connect to local MySQL server through socket '/var/run/mysqld/mysqld.sock' (2) on subsequent connection through terminal
So basically you need to tweak the [client] section of your my.cnf to set up the correct location of mysqld.sock (it should be /tmp/mysqld.sock I guess). You can't just copy the socket file because it will be recreated in the old location on MySQL restart.