Skip to contentSkip to conversation

hanging in wait_for_tcp_memory

DISCUSSION Gerald Krafft 2 participants 3 messages Mar 11, 2004

lore
Share & tools
lore
#0 lore raw

I have some database processes (Interbase V6) that occasionally seem to hang. Using ps or top I found that they are waiting in wait_for_tcp_memory. What exactly does wait_for_tcp_memory do and under which circumstances does this function block? I'm using Red Hat Linux 7.2, kernel version 2.4.7-10smp on a dual processor machine. Were there any known problems with wait_for_tcp_memory in that kernel version that might have been fixed in later versions?

Thanks
Gerald

#1 lore raw
Quoted 8 lines · Gerald Krafft

I have some database processes (Interbase V6) that occasionally seem to hang. Using ps or top I found that they are waiting in wait_for_tcp_memory. What exactly does wait_for_tcp_memory do and under which circumstances does this function block?
I'm using Red Hat Linux 7.2, kernel version 2.4.7-10smp on a dual processor machine. Were there any known problems with wait_for_tcp_memory in that kernel version that might have been fixed in later versions

2.4.7-10 had a really really bad vm, so we replaced it with another kernel the day of 7.2 release (which also was needed for security fixes). Sounds like you need to apply a bunch of (security) errata to get your system performing better...

#2 lore raw

Thanks, so I will have to upgrade my system to a later kernel. It would still be nice if somebody could explain what this kernal function does, when it is usually called and under which circumstances it might block.

Gerald

Quoted 10 lines · Arjan van de Ven
Quoted 2 lines · Gerald Krafft

I have some database processes (Interbase V6) that occasionally seem to hang. Using ps or top I found that they are waiting in wait_for_tcp_memory. What exactly does wait_for_tcp_memory do and under which circumstances does this function block? I'm using Red Hat Linux 7.2, kernel version 2.4.7-10smp on a dual processor machine. Were there any known problems with wait_for_tcp_memory in that kernel version that might have been fixed in later versions

2.4.7-10 had a really really bad vm, so we replaced it with another kernel the day of 7.2 release (which also was needed for security fixes). Sounds like you need to apply a bunch of (security) errata to get your system performing better...