我已经安装了 CouchDB v1.0.4,我想用 Supervisor 监控它的进程(安装的版本是 2.1.3)。我可以用以下命令启动 CouchDB:
$ /usr/bin/couchdb
我对其有以下基本配置/etc/supervisord.d
:
[program:couchdb]
command=/usr/bin/couchdb
但是当我启动时supervisord
它不会couchdb
按预期启动。错误日志/var/log/supervisor/supervisord.log
只是确认启动后该进程存在:
2014-09-25 13:34:10,990 INFO /var/tmp/supervisor.sock:Medusa (V1.1.1.1) started at Thu Sep 25 13:34:10 2014
Hostname: <unix domain socket>
Port:/var/tmp/supervisor.sock
2014-09-25 13:34:11,024 CRIT Running without any HTTP authentication checking
2014-09-25 13:34:11,024 INFO daemonizing the process
2014-09-25 13:34:11,025 INFO supervisord started with pid 8258
2014-09-25 13:34:11,037 INFO spawned: 'couchdb' with pid 8259
2014-09-25 13:34:11,044 INFO exited: couchdb (exit status 1; not expected)
2014-09-25 13:34:11,044 INFO received SIGCLD indicating a child quit
2014-09-25 13:34:12,051 INFO spawned: 'couchdb' with pid 8269
2014-09-25 13:34:12,274 INFO exited: couchdb (exit status 1; not expected)
2014-09-25 13:34:12,274 INFO received SIGCLD indicating a child quit
2014-09-25 13:34:14,299 INFO spawned: 'couchdb' with pid 8297
2014-09-25 13:34:14,300 INFO exited: couchdb (exit status 1; not expected)
2014-09-25 13:34:14,300 INFO received SIGCLD indicating a child quit
2014-09-25 13:34:18,319 INFO spawned: 'couchdb' with pid 8306
2014-09-25 13:34:18,320 INFO exited: couchdb (exit status 1; not expected)
2014-09-25 13:34:18,320 INFO received SIGCLD indicating a child quit
2014-09-25 13:34:19,320 INFO gave up: couchdb entered FATAL state, too many start retries too quickly
我真的不知道为什么会这样,因为没有记录任何错误/var/log/couchdb/couch.log
(CouchDB 的日志级别设置为debug
)。
任何帮助将不胜感激。
答案1
你必须在supervisord配置中为couchdb设置环境$HOME
environment=HOME=/usr/local/var/lib/couchdb redirect_stderr=true