Menu

#4581 conserver always configures for 1 hour timestamps. we need timestamps per line.

2.11
open
None
hw ctrl
5
2015-07-06
2015-02-27
No

The IBM CORAL project requires that we take the console log and coorelate events in it along with all other logs in the system. this requires timestamps on every line.

Looking in conserver.pm in the xcat source I see this hard coded:

push @newheaders,"default * {\n";
push @newheaders," logfile /var/log/consoles/&;\n";
push @newheaders," timestamp 1hab;\n";
push @newheaders," rw *;\n";
push @newheaders," master localhost;\n";

We need the ability to be able to control the contents of the timestamp value and change it to "timestamp 1l;".

Is it possible to add this is a feature for future releases.

Discussion

  • Guang Cheng Li

    Guang Cheng Li - 2015-03-02
    • assigned_to: Victor Hu
    • component: unknown --> hw ctrl
     
  • Guang Cheng Li

    Guang Cheng Li - 2015-03-02

    We are replacing conserver with a xCAT-developed tool named confluent, but I think this is a reasonable requirement for both conserver based console and confluent based console.

    Victor, could you help on this bug? Thanks.

    I did some quick experiment, changing the timestamp line in /etc/conserver.cf to be "timestamp 1lab" does trigger the conserver to log timestamp for each line.

    [root@ls21n01 consoles]# cat /etc/conserver.cf | grep timestamp
      timestamp 1lab;
    [root@ls21n01 consoles]# tail -f /var/log/consoles/rhcn2
    ...
    [Sun Mar  1 20:18:41 2015]/dev/sda3: clean, 31181/533280 files, 285292/2131456 blocks^M
    [Sun Mar  1 20:18:41 2015][/sbin/fsck.ext3 (1) -- /boot] fsck.ext3 -a /dev/sda1 ^M
    [Sun Mar  1 20:18:41 2015]/dev/sda1: recovering journal^M
    [Sun Mar  1 20:18:42 2015]/dev/sda1: clean, 39/65536 files, 46957/262144 blocks^M
    [Sun Mar  1 20:18:42 2015][  OK  ]^M
    [Sun Mar  1 20:18:42 2015]Remounting root filesystem in read-write mode:  [  OK  ]^M
    [Sun Mar  1 20:18:42 2015]Mounting local filesystems:  kjournald starting.  Commit interval 5 seconds^M
    [Sun Mar  1 20:18:42 2015]EXT3-fs (sda1): using internal journal^M
    [Sun Mar  1 20:18:42 2015]EXT3-fs (sda1): mounted filesystem with ordered data mode^M
    [Sun Mar  1 20:18:42 2015][  OK  ]^M
    [Sun Mar  1 20:18:42 2015]Enabling local filesystem quotas:  [  OK  ]^M
    [Sun Mar  1 20:18:42 2015]Enabling /etc/fstab swaps:  Adding 975864k swap on /dev/sda2.  Priority:-1 exten^M
    [Sun Mar  1 20:18:42 2015]^[[79C^[[A^@^@^@^@^@^@^@^@nts:1 across:975864k ^M
    [Sun Mar  1 20:18:42 2015][  OK  ]^M
    [Sun Mar  1 20:18:42 2015]Entering non-interactive startup^M
    [Sun Mar  1 20:18:42 2015]Calling the system activity data collector (sadc)... ^M
    [Sun Mar  1 20:18:42 2015]NET: Registered protocol family 10^M
    [Sun Mar  1 20:18:42 2015]lo: Disabled Privacy Extensions^M
    [Sun Mar  1 20:18:42 2015]Bringing up loopback interface:  [  OK  ]^M
    [Sun Mar  1 20:18:43 2015]Bringing up interface eth0:  ^M
    [Sun Mar  1 20:18:43 2015]Determining IP information for eth0... done.^M
    [Sun Mar  1 20:18:46 2015][  OK  ]^M
    ....
    
     
  • Guang Cheng Li

    Guang Cheng Li - 2015-03-02

    We might need to evaluate if this will bring in any performance overhead by enabling the per-line timestamp in the console log.

     
  • Guang Cheng Li

    Guang Cheng Li - 2015-07-06
    • Milestone: 2.10 --> 2.11
     
  • Guang Cheng Li

    Guang Cheng Li - 2015-07-06

    Needs to verify replaycons also works with the per line timestamp

     
  • ralph bellofatto

    I am out of the office until 07/20/2015.

    I will respond to your message when I return.

    Note: This is an automated response to your message "[xcat:bugs] #4581
    conserver always configures for 1 hour timestamps. we need timestamps per
    line." sent on 07/06/2015 4:45:47 AM.

    This is the only notification you will receive while this person is away.

     
MongoDB Logo MongoDB