我在前台运行服务.有时它会被终止,大概是出于内存的原因(尽管我对LogCat日志并不是100%肯定).已终止的其他服务计划在5000毫秒内重新启动,但查看日志,我的服务重新启动时间很长,例如:
- 11-15 15:39:48.756: W/ActivityManager(375): Scheduling restart of crashed service com.example.app/com.example.SensorService in 1019562ms
解决方法
从
com/android/server/am/ActiveServices.java(未曝光)
- // How long we wait for a service to finish executing.
- static final int SERVICE_TIMEOUT = 20*1000;
- // How long a service needs to be running until restarting its process
- // is no longer considered to be a relaunch of the service.
- static final int SERVICE_RESTART_DURATION = 5*1000;
- // How long a service needs to be running until it will start back at
- // SERVICE_RESTART_DURATION after being killed.
- static final int SERVICE_RESET_RUN_DURATION = 60*1000;
- // Multiplying factor to increase restart duration time by,for each time
- // a service is killed before it has run for SERVICE_RESET_RUN_DURATION.
- static final int SERVICE_RESTART_DURATION_FACTOR = 4;
- // The minimum amount of time between restarting services that we allow.
- // That is,when multiple services are restarting,we won't allow each
- // to restart less than this amount of time from the last one.
- static final int SERVICE_MIN_RESTART_TIME_BETWEEN = 10*1000;
- // Maximum amount of time for there to be no activity on a service before
- // we consider it non-essential and allow its process to go on the
- // LRU background list.
- static final int MAX_SERVICE_INACTIVITY = 30*60*1000;
你发生的事情可能是你的服务比SERVICE_RESET_RUN_DURATION快死,然后重启时间乘以SERVICE_RESTART_DURATION_FACTOR.
从第881行开始:
- // If it has been a "reasonably long time" since the service
- // was started,then reset our restart duration back to
- // the beginning,so we don't infinitely increase the duration
- // on a service that just occasionally gets killed (which is
- // a normal case,due to process being killed to reclaim memory).
- if (now > (r.restartTime+resetTime)) {
- r.restartCount = 1;
- r.restartDelay = minDuration;
- } else {
- if ((r.serviceInfo.applicationInfo.flags
- &ApplicationInfo.FLAG_PERSISTENT) != 0) {
- // Services in peristent processes will restart much more
- // quickly,since they are pretty important. (Think systemUI).
- r.restartDelay += minDuration/2;
- } else {
- r.restartDelay *= SERVICE_RESTART_DURATION_FACTOR;
- if (r.restartDelay < minDuration) {
- r.restartDelay = minDuration;
- }
- }
- }
下面的行将支持您的服务.
- r.restartDelay *= SERVICE_RESTART_DURATION_FACTOR;
因此,如果您的服务在运行SERVICE_RESET_RUN_DURATION之前死亡,您应该修复此案例.