ArangoDB alway restart when Arangodb change log file.

[liuxiong888]

My Environment

• __ArangoDB Version: 3.6.3-1
• Storage Engine: RocksDB
• Deployment Mode: Single Server
• Deployment Strategy: Manual Start
• Configuration:
• Infrastructure:
• Operating System: CentOS7
• Total RAM in your machine: 16GB
• Disks in use: 100GB
• Used Package:

Component, Query & Data

Size of your Dataset on disk: 15GB

Steps to reproduce

1. Look the log for detal

Problem:
ArangoDB alway restart when Arangodb change log file by himself.

Expected result: Stop auot restart.

[liuxiong888]

[
arangod.log
](url)

[liuxiong888]

There is also another restart:
2020-05-28T09:02:50Z [7200] INFO [e52b0] ArangoDB 3.6.3-1 [linux] 64bit, using jemalloc, build tags/v3.6.3.1-0-g76b2e00b64b-dirty, VPack 0.1.33, RocksDB 6.2.0, ICU 58.1, V8 7.1.302.28, OpenSSL 1.1.1g 21 Apr 2020
2020-05-28T09:02:50Z [7200] INFO [75ddc] detected operating system: Linux version 3.10.0-1062.9.1.el7.x86_64 (mockbuild@kbuilder.bsys.centos.org) (gcc version 4.8.5 20150623 (Red Hat 4.8.5-39) (GCC) ) #1 SMP Fri Dec 6 15:49:49 UTC 2019
2020-05-28T09:02:50Z [7200] INFO [25362] {memory} Available physical memory: 16656625664
2020-05-28T09:02:50Z [7200] INFO [43396] {authentication} Jwt secret not specified, generating...
2020-05-28T09:02:50Z [7200] INFO [144fe] using storage engine 'rocksdb'
2020-05-28T09:02:50Z [7200] INFO [3bb7d] {cluster} Starting up with role SINGLE
2020-05-28T09:02:50Z [7200] INFO [6ea38] using endpoint 'http+tcp://127.0.0.1:8529' for non-encrypted requests
2020-05-28T09:02:50Z [7200] INFO [6ea38] using endpoint 'http+tcp://172.27.0.9:8529' for non-encrypted requests
2020-05-28T09:02:50Z [7200] INFO [a1c60] {syscall} file-descriptors (nofiles) hard limit is 131072, soft limit is 131072
2020-05-28T09:02:50Z [7200] INFO [3844e] {authentication} Authentication is turned on (system only), authentication for unix sockets is turned on
2020-05-28T09:02:50Z [7200] WARNING [b387d] found existing lockfile '/var/lib/arangodb3/LOCK' of previous process with pid 2118, but that process seems to be dead already
2020-05-28T09:02:58Z [7200] INFO [cf3f4] ArangoDB (version 3.6.3-1 [linux]) is ready for business. Have fun!

[liuxiong888]

Anyone can tell me what's this meaning and how to resolve it :file-descriptors (nofiles) hard limit is 131072, soft limit is 131072

[jsteemann]

I am having trouble to understand the problem...
Are you saying you are rotating the logfile and then ArangoDB restarts? What exact action are you carrying out that causes ArangoDB to restart?
Thanks!

[jsteemann]

Anyone can tell me what's this meaning and how to resolve it :file-descriptors (nofiles) hard limit is 131072, soft limit is 131072

This is just an informational message. It does not indicate a problem. It shows the number of available file descriptors available for the ArangoDB process. A low value (say 1024) would be problematic, but 131072 seems more than enough.
So all good.

[liuxiong888]

About every week,Arangodb will create a new logfile and backup old one. But the ArangoDB also restart then.

[dothebart]

Hi,
ArangoDB supports re-opening its log-files using the SIG_HUP signal - as it is unix standard:
https://unix.stackexchange.com/questions/440004/why-is-kill-hup-used-in-logrotate-in-rhel-is-it-necessary-in-all-cases
whats your logrotate configuration? Is it similar to the one in the SO-question?
Moving the files should be done and controlled by the system service named logrotate.

[Simran-B]

@liuxiong888 Can you provide for information as requested by dothebart?