Forum Discussion

Nick_J's avatar
Nick_J
Level 4
11 years ago

Media server is not being 'active for tape and disk jobs' intermittently

Media server: Runs on Linux 2.6.18-128.el5 and with NBU version 7.5.0.5   The media server state to 'active for tape and disk jobs' is intermittently going off to 'active for disk jobs'. So far I...
7.5
Backup and Recovery
Best Practice
Error messages
Linux
NetBackup
Tip-How to
Troubleshooting
  • Nick_J's avatar
    Nick_J
    11 years ago

    Observed the following errors in the media server.

    /var/log/messages:

    tldd[26336]: Error in Semaphore Operation - Invalid argument
    tldd[26336]: Daemon has terminated abnormally
    tldd[26336]: Daemon has terminated due to IPC message failure
    ntpd[1227]: sendto(xx.xx.xx.xx) (fd=-1): Bad file descriptor
    vmd[27325]: terminating - another daemon already exists (89)
     
    Daemon log:
     
    <16> no_daemon_exists: file already locked
    <16> vmd: vmd: Unable to get daemon lock file
    <16> vmd: terminating - another daemon already exists (89)
     
    Deleted the configured devices using the following command and reconfigured the tape drives from the scratch.
    nbemmcmd -deletealldevices -machinetype media -machinename xxxx
     
    tpconfig -dev_ping
    Confirmed the media server could be able to talk with the tape drives:
    tpconfig -dev_ping -drive -path /dev/xxxx
     
    I strongly believed the hung/stuck semaphore in the OS and unexpected terminated lock files, along with the large operation of bad descriptors were dominating this issue, hence bounced the media server, from then on the MachineState is stable for tape and disk jobs.

     

Nick_J's avatar
Nick_J
Level 4
11 years ago

Observed the following errors in the media server.

/var/log/messages:

tldd[26336]: Error in Semaphore Operation - Invalid argument
tldd[26336]: Daemon has terminated abnormally
tldd[26336]: Daemon has terminated due to IPC message failure
ntpd[1227]: sendto(xx.xx.xx.xx) (fd=-1): Bad file descriptor
vmd[27325]: terminating - another daemon already exists (89)
 
Daemon log:
 
<16> no_daemon_exists: file already locked
<16> vmd: vmd: Unable to get daemon lock file
<16> vmd: terminating - another daemon already exists (89)
 
Deleted the configured devices using the following command and reconfigured the tape drives from the scratch.
nbemmcmd -deletealldevices -machinetype media -machinename xxxx
 
tpconfig -dev_ping
Confirmed the media server could be able to talk with the tape drives:
tpconfig -dev_ping -drive -path /dev/xxxx
 
I strongly believed the hung/stuck semaphore in the OS and unexpected terminated lock files, along with the large operation of bad descriptors were dominating this issue, hence bounced the media server, from then on the MachineState is stable for tape and disk jobs.