-
Notifications
You must be signed in to change notification settings - Fork 41
jvm.options: Error opening log file 'logs/gc.log': No such file or directory #249
Comments
Would be nice if you guys could fix this pronto... somewhat unprofessional to download the software, follow the steps it tells you, and then it doesn't work. Yes, I know the MSI is beta, but a modicum of testing would have been nice. |
Sorry about the delay in updating here. Our CI runs against Java 8 and does not surface this error. Will look into getting this fixed again this week. |
I have same issue to install and run elastic search on Windows 10.
|
I have same issue on Windows 7. OpenJDK 11.0.2 and ES 6.6.0. Used .msi to install.
|
I resolved this on Windows 10 as follows:
|
Can confirm the steps given above to replicate the issue:
Currently working on a fix; new installations will modify the |
Thank you guys! |
I face with the same problem in windows, I create the |
@lugq1990 This has been fixed in 7.2.1 and 7.3.0. Would you be able to try with one of those versions? |
I have the same issue..But in mac osX `Macintosh:~ sura_developers$ elasticsearch [2019-08-20T16:18:11,493][WARN ][o.e.b.ElasticsearchUncaughtExceptionHandler] [unknown] uncaught exception in thread [main] |
Please help me to solve this problem. |
@nijanthan0 I had this issue with elasticsearch installed via brew on mac Mojave and was able to resolve it with the following steps:
This is for elasticsearch 6.8.1 installed via homebrew with the following output from
|
@vaelinn123 @nijanthan0
|
make a gc.log life inside /usr/local/var/log/elasticsearch/ using touch gc.log 8:-Xloggc:/usr/local/var/log/elasticsearch/gc.log JDK 9+ GC logging9-:-Xlog:gc*,gc+age=trace,safepoint:file=/usr/local/var/log/elasticsearch/gc.log:utctime,pid,tags:filecount=32,filesize=64m and restart services |
for me, there was an error, because previusly using Homebrew, i install version 6.8.3. To solve, i uninstall 7.6.0 first, then remove |
Environment
Server 2012 R2 Standard
Oracle Java 10.0.1+10, Java HotSpot(TM) 64-Bit Server VM
Steps to reproduce the behaviour
New install of
elasticsearch-6.5.2.msi
. I noticed the issue when using vagrant + ansible to provision the system. I uninstalled, and reinstalled using the UI (and mainly defaults, basic license, and untick start server [as I believe this causes the install to fail, since the service can't start]). Starting elasticsearch manually through powershell (or command prompt), I get the following error:If I edit the
C:\ProgramData\Elastic\Elasticsearch\config\jvm.options
file, and change line105
fromsafepoint:file=logs/gc.log
->safepoint:file=/ProgramData/Elastic/Elasticsearch/log/gc.log
I can successfully start elasticsearch.I suspect that this location should actually resolve to
%ALLUSERSPROFILE%/Elastic/Elasticsearch/log
, since that is where other elasticsearch log will go to. The jvm needs to know the right path in this case, since it can't read the log location from the elasticsearch.yml until the process has already started.There are other relative paths in the jvm.options as well:
8:-Xloggc:logs/gc.log
This would likely cause the same issue as described above, if I was using Java 8. Since I am not, this seems to be ignored and elasticsearch can start.-XX:ErrorFile=logs/hs_err_pid%p.log
. This doesn't prevent elasticsearch from starting, but likely means that a heap dump on OOM will fail due to the relative path.related issues:
cc: @Mpdreamz
The text was updated successfully, but these errors were encountered: