Seafile high java CPU load


i am using seafile pro 7.0.7 on a debian buster root server and started to notice something that seems suspicious to me. When i start htop there are several java entries that repeatedly seem to consume lots of cpu for a few seconds. After that the cpu load seems perfectly normal until it comes up back again like you see in the picture.

Even though there are no obvious problems when we are using seafile so far i still would like to know
if this is intented or something i should look into more closely.

Thanks in advance for your help



actual I have the same problem. Maybe since I upgrade from 6.x to 7.x.
I hope anybody can help us. :slight_smile:


I disabled the file index by setting

enabled = false

in ./conf/seafevents.conf. The cpu and memory load is normal again.

Elasticsearch seems to be the problem. I found this line in the elasticssearch.log:

[2019-09-03T15:37:53,897][WARN ][o.e.b.ElasticsearchUncaughtExceptionHandler] [] uncaught exception in thread [main]
org.elasticsearch.bootstrap.StartupException: java.lang.IllegalStateException: failed to obtain node locks, tried [[/opt/srv/seafile/pro-data/search/data/elasticsearch]] with lock id [0]; maybe these locations are not writable or multiple nodes were started without increasing [node.max_local_storage_nodes] (was [1])?
OK… it on warn level, but elasticsearch will be restarted and restarted again and again. I guess this is the reason for the high system load.

so I deleted the entire ‘pro-data/search’ folder and activated the file index and started seafile again. I hoped the file index would be created and all will be fine but this didn’t fixed my problem. CPU load is on max.

I deactivated file indexing again and I’m waiting for a solution.

I can confirm what you described. Immediately after setting INDEX FILES to false, the high CPU load disappears completly. You might also be able to install elasticsearch seperately and set the
es_host and and es_ports setting in seafevents.conf accordingly.

If any of the devs is seeing this, it would be good to know if this can be considered a bug and if they are fixing it.