Closed Si-Richards closed 4 years ago
My understanding is that restarting Heplify server would drop old data but this is not happening...
Not sure where this is determined from. HEPlify-server needs to be running for rotation to take place and should remove data once it surpasses the configured age. Please share your config and results vs. expected results
Heplify is running with the config above, even if i restart the service there is no drop.
My heplify log is:
2019-10-01T11:41:44+01:00 WARN pq: the database system is starting up 2019-10-01T11:43:36+01:00 INFO end creating tables (2019-10-01 11:43:36.991474764 +0100 BST m=+113.046494238) 2019-10-01T11:43:39+01:00 INFO postgres connection established 2019-10-01T11:43:39+01:00 ERR pq: could not extend file "base/16385/74378": No space left on device 2019-10-01T11:43:39+01:00 ERR pq: Could not complete operation in a failed transac
/opt/heplify-server# df Filesystem 1K-blocks Used Available Use% Mounted on devtmpfs 4903008 0 4903008 0% /dev tmpfs 4914756 0 4914756 0% /dev/shm tmpfs 4914756 8708 4906048 1% /run tmpfs 4914756 0 4914756 0% /sys/fs/cgroup /dev/mapper/centos-root 36678148 36678128 20 100% / /dev/sda1 1038336 236632 801704 23% /boot tmpfs 982952 0 982952 0% /run/user/0
Hi @Si-Richards , I see that you use the config option DBDropOnStart = true Indeed this means that tables older then DBDropDays = 1 should be deleted when you start heplify-server.
But from the latest log I see ERR pq: could not extend file "base/16385/74378": No space left on device It seems that you don't have room for vacuuming and thats why heplify-server fails to drop old partitions.
@Si-Richards were you able to solve this so I can close this issue?
No response->close.
@Si-Richards were you able to solve this so I can close this issue?
Sorry for the lack of reply, we moved to Homer Legacy as it was not solved.
I've setup Homer7 again and still get the issue.....
2019-12-03T09:49:33Z INFO start heplify-server 1.12 with config.HeplifyServer{HEPAddr:"0.0.0.0:9060", HEPTCPAddr:"", HEPTLSAddr:"0.0.0.0:9060", CGRAddr:"", ESAddr:"", ESDiscovery:true, ESUser:"", ESPass:"", LokiURL:"", LokiBulk:200, LokiTimer:4, LokiBuffer:100000, LokiHEPFilter:[]int{}, ForceHEPPayload:[]int{}, PromAddr:"0.0.0.0:9096", PromTargetIP:"", PromTargetName:"", DBShema:"homer7", DBDriver:"postgres", DBAddr:"localhost:5432", DBUser:"homer", DBPass:"DMK7VVhhweYANCCUvYXOwPaj6gI", DBDataTable:"homer_data", DBConfTable:"homer_config", DBBulk:200, DBTimer:4, DBBuffer:400000, DBWorker:8, DBRotate:true, DBPartLog:"2h", DBPartIsup:"6h", DBPartSip:"1h", DBPartQos:"6h", DBDropDays:1, DBDropDaysCall:1, DBDropDaysRegister:1, DBDropDaysDefault:1, DBDropOnStart:true, Dedup:false, DiscardMethod:[]string{}, AlegIDs:[]string{}, CustomHeader:[]string{}, LogDbg:"", LogLvl:"info", LogStd:false, LogSys:false, Config:"./heplify-server.toml", ConfigHTTPAddr:"", ConfigHTTPPW:""}
2019-12-03T09:49:33Z INFO expose metrics without or unbalanced targets 2019-12-03T09:49:33Z INFO start creating tables (2019-12-03 09:49:33.162173299 +0000 GMT m=+0.012348456)
2019-12-03T09:49:33Z WARN dial tcp [::1]:5432: connect: connection refused 2019-12-03T09:49:33Z WARN dial tcp [::1]:5432: connect: connection refused
Postgres is unreachable based on the error message alone. How are you setting up the system?
Hi Imangani,
This time i am using your script for centos7 (https://github.com/sipcapture/homer-installer/tree/7.7) It's strange because yes i can see that the partition is 100% from df -h. If i restart posgres and heplify services then Homer has completed a clean up and i am now only 28% used but i had to force this with the restart.
BTW, the UI looks really good now!
so as a hackey work around i have a cronjob to restart the services each day.
--- cut --- DBUser:"homer", DBPass:"DMK7VVhhweYANCCUvYXOwPaj6gI", DBDataTable:"homer_data", DBConfTable:"homer_config" --- cut ---
@Si-Richards I suggest you change the database password too.. 😇
--- cut --- DBUser:"homer", DBPass:"DMK7VVhhweYANCCUvYXOwPaj6gI", DBDataTable:"homer_data", DBConfTable:"homer_config" --- cut ---
@Si-Richards I suggest you change the database password too.. 😇
Thats a fair point......
As a side note i looked at the rotator code as i really need to rotate after 12hrs as i'm collecting over 50Gb a day.
Shame there isnt a way to force a shorter rotate
Feel free to do a PR and when you are doing it please feel free to add some tests for it too. Thanks!
Btw. my gradma is using usb sticks bigger than 50GB :D
So what I want to say is that I see no sense in implementing this myself. I mean 100GB a day what is this? 3TB a month, common thats nothing ;)
I have homer7 on Centos 7 and the PostgreSQL 10 database server is not rotating. My understanding is that restarting Heplify server would drop old data but this is not happening...
My config looks like this: