Project

General

Profile

Actions

tickets #39173

closed

keyserver1: /srv/sks nearly full

Added by cboltz over 5 years ago. Updated over 5 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Servers hosted in NBG
Target version:
-
Start date:
2018-08-05
Due date:
% Done:

0%

Estimated time:

Description

Monitoring says that the /srv/sks partition on keyserver1 is nearly full.

The reason is probably that there are lots of /srv/sks/KDB/log.0000* files:

keyserver1 (keyserver1.o.o):/srv/sks/KDB # du -h log.000000* --total | tail -n1
78G     total

Please handle this ASAP - ideally before the partition is really full (currently only 200 MB free!), and something breaks.

I have no idea if these log files are still needed, therefore I won't touch them ;-)

Actions #1

Updated by cboltz over 5 years ago

  • Private changed from Yes to No
Actions #2

Updated by TBro over 5 years ago

  • Status changed from New to In Progress
  • Assignee changed from tampakrap to TBro

working on it.
Solution for next time.

There are two approaches:
1) (automated) Add this line to DB_CONFIG (which you are using, correct?)
set_flags DB_LOG_AUTOREMOVE
2) (manual) Run these commands on idle databases (with KDB/PTree as cwd)
dbXY_recover -ev
dbXY_checkpoint -1
dbXY_archive -dv
(XY == version of Berkeley DB in use, e.g. XY=53 for db-5.3.x)

Needs to be done in the directory: /srv/sks/KDB

Actions #3

Updated by TBro over 5 years ago

fixed for now by placing DB_CONFIG file from here:
https://bitbucket.org/skskeyserver/sks-keyserver/src/0106ba2f6abc8ee2f685026d3f218730114a60fa/sampleConfig/DB_CONFIG?at=default&fileviewer=file-view-default

manually to /etc/sks and restarted all services

Opened a Bug as well - as I don't have time now to look into this detailed: https://bugzilla.opensuse.org/show_bug.cgi?id=1103862

Actions #4

Updated by TBro over 5 years ago

  • Status changed from In Progress to Resolved
Actions

Also available in: Atom PDF