Filter
Top Products
Alarms in Linux Media Servers
Issue 1 June 2005 105
13
(Except
S8500)
MIN “ERROR: could not dup socket fd in reqsvr.c, heartbeat thread not created,
errno=<x>” — As either a:
● Booting application initiates
● Restarting application re-initiates heart beating with Watchdog (see
Event ID #5
), the “reqsvr” (request server) thread tries to create a
duplicate socket for the hearbeating thread. This alarm indicates that
reqsvr could not create the socket. (Meanwhile, if the other server is
already started or subsequently starts, it assumes the active role.)
In that (when this alarm occurs) a server hangs, this alarm’s external
symptoms resemble those of two other _WD alarms, #5
and #14
(Except S8500). Therefore, carefully discriminate between these three
events.
1. To inspect the symptoms of this problem, verify that the:
a. Linux OS and the Web interface are up (including the commands:
telnet, statapp, server, logv, cat, grep, tail, vi, etc.)
b. Watchdog application is partially up, and some other
Communication Manager processes may be up, either using the:
- Web interface, by selecting View Process Status
- Linux command line, by entering statapp
c. If the GMM process is up, the almdisplay -v command shows
Event #13’s message string.
If not, the Watchdog log shows the message string, either using
the:
- Web interface, by selecting Diagnostics > View System Logs
and Watchdog Logs
- Linux command line, by entering logv -w or, directly, by
examining /var/log/ecs/wdlog.
Since the following commands cause a brief service outage, they should
only be executed at the customer’s convenience.
2. Escalate this problem for explicit guidance with the following steps.
3. Enter server to verify that the suspected server is the standby.
If necessary and at the customer’s convenience, enter server -if
to force a server interchange.
Table 31: _WD Alarms in Media Server (continued)
Event
ID
Alarm
Level
Alarm Text, Cause/Description, Recommendation
6 of 15