eleaner
eleaner
Thank you you are a life savier On Sun, 2 Dec 2018 at 01:21, Chai Feng wrote: > I found that the script doesn't clean up the previous rule for...
I am pretty sure I followed the best practices including number 7. vm.overcommit_memory is set to 1 The only exception is the hdd disk space - I have allocated half...
ok, got another crash I cannot find anything unusual in dmesg master log ``` Aug 06 19:00:03 mercury mfsmaster[7249]: created new sessionid:13392 Aug 06 19:00:03 mercury mfsmaster[7249]: created new sessionid:13393...
master has 16GB of RAM + 8GB swap max usage recorded by zabbix = almost 12GB master currently uses 5.3 GiB
another crash, this time I missing metadata.mfs afterwards that's definitely not good, any ideas? ``` Aug 08 17:00:10 mercury mfsmaster[28326]: child finished Aug 08 17:00:10 mercury mfsmaster[28326]: store process has...
I removed moosefs-master git cloned the repository run linux_build.sh then it stopped with "zlib development library not found" so I stopped here and reinstalled the 3.0.105 I cannot afford for...
@oxide94 It crashed again after I add a new chunkserver, so it might be related to high I/O since my master is also a chunk server. Unfortunately, I did not...
and now it crashes every few minutes with the same 'connection lost with background data writer - exiting' my dmesg is full of messages like that but I am not...
@acid-maker not that I can see. no packages from @oxide94
Running 3.0.106 second day stable