Re: [GENERAL] Not your father's question about deadlocks

2006-11-17 Thread Tom Lane
Clarence Gardner <[EMAIL PROTECTED]> writes: > I'm still creeping up on the problem. I configured postgres to resolve > a deadlock only after a month, and now I seem to have a complete > snapshot of one of them. If you don't mind rebuilding Postgres from source, here's the patch as applied to 8.1

Re: [HACKERS] [GENERAL] Not your father's question about deadlocks

2006-11-16 Thread Gurjeet Singh
On 11/17/06, Tom Lane <[EMAIL PROTECTED]> wrote: "Gurjeet Singh" <[EMAIL PROTECTED]> writes: > Small implementation detail: Also keep a count of how many times the same > session requested the same lock, and do not release the lock until he > requests same number of releases. No need for that,

Re: [HACKERS] [GENERAL] Not your father's question about deadlocks

2006-11-16 Thread Tom Lane
"Gurjeet Singh" <[EMAIL PROTECTED]> writes: > On 11/17/06, Tom Lane <[EMAIL PROTECTED]> wrote: >> we need a special case when we are already a member of the MultiXact: >> fall through without trying to reacquire the tuple lock. > Small implementation detail: Also keep a count of how many times the

Re: [HACKERS] [GENERAL] Not your father's question about deadlocks

2006-11-16 Thread Gurjeet Singh
On 11/17/06, Tom Lane <[EMAIL PROTECTED]> wrote: we need a special case when we are already a member of the MultiXact: fall through without trying to reacquire the tuple lock. Small implementation detail: Also keep a count of how many times the same session requested the same lock, and do not

Re: [GENERAL] Not your father's question about deadlocks

2006-11-16 Thread Tom Lane
Clarence Gardner <[EMAIL PROTECTED]> writes: > That scenario seems quite simple, but I can't reproduce the deadlock with > this seemingly-identical sequence. This is a bug in 8.1 and up. The reason you couldn't reproduce it is that it requires a minimum of three transactions involved, two of whic

[GENERAL] Not your father's question about deadlocks

2006-11-15 Thread Clarence Gardner
Once upon a time, I put a question regarding deadlocks to the group, and Tom Lane immediately answered with this: >The guy waiting on the tuple-specific lock is second in >line to actually mung the tuple. Whoever is first in line behind the >current tenant will be blocked trying to acquire Share

[GENERAL] Not your father's question about deadlocks

2006-11-14 Thread Clarence Gardner
Once upon a time, I put a question regarding deadlocks to the group, and Tom Lane immediately answered with this: >The guy waiting on the tuple-specific lock is second in >line to actually mung the tuple. Whoever is first in line behind the >current tenant will be blocked trying to acquire ShareL