.
MGC ELAN goes down and gateway appears to be locked up after BUG7058 IN cemux task manual reboot required to recover gateway.
MGC elan goes down on call server and does not recover.
MGC logs indicate BUG7058 SWD in CEMUX tas
The problem was that the MGC locked up. The MGC was not at all responsive from a network perspective - the ELAN link to the CS was down. No dial tone from any of the sets. Could not telnet or rlogin to the card. the card was power down rebooted to recover. From looking at the logs the cause the only evidence on the MGC of any sort of problem was 4 BUG7058's. Task Timeout for the Cemux tasks
The cause is that each of these tasks were pending on the SPLNET semaphore . The SPLNET semaphore is a system wide network semaphore. Any task that needs any type of network access requires this semaphore. Since neither of these two tasks could get the semaphore for 1000 seconds it is safe to assume no task could. This matches the symptoms perfectly - no network access. If the site was rebooted with a reboot-1 i would already have a solution
Solution:
MGC patch developed MPLR32222
Patch will do 2 things. Please only use it at this specific site.
1. When we determine the cemux tasks are timed out we will print the owner of the
splnet semaphore.
2. reboot the system to automatically recover
MGC elan goes down on call server and does not recover.
MGC logs indicate BUG7058 SWD in CEMUX tas
The problem was that the MGC locked up. The MGC was not at all responsive from a network perspective - the ELAN link to the CS was down. No dial tone from any of the sets. Could not telnet or rlogin to the card. the card was power down rebooted to recover. From looking at the logs the cause the only evidence on the MGC of any sort of problem was 4 BUG7058's. Task Timeout for the Cemux tasks
The cause is that each of these tasks were pending on the SPLNET semaphore . The SPLNET semaphore is a system wide network semaphore. Any task that needs any type of network access requires this semaphore. Since neither of these two tasks could get the semaphore for 1000 seconds it is safe to assume no task could. This matches the symptoms perfectly - no network access. If the site was rebooted with a reboot-1 i would already have a solution
Solution:
MGC patch developed MPLR32222
Patch will do 2 things. Please only use it at this specific site.
1. When we determine the cemux tasks are timed out we will print the owner of the
splnet semaphore.
2. reboot the system to automatically recover
