[prev in list] [next in list] [prev in thread] [next in thread] 

List:       squid-dev
Subject:    Re: diskd assertion
From:       Duane Wessels <wessels () ircache ! net>
Date:       2000-05-30 4:12:27
Message-ID: Pine.SGI.4.10.10005292208370.17756-100000 () surf ! ircache ! net
[Download RAW message or body]



On Tue, 30 May 2000, Henrik Nordstrom wrote:

> Can someone care to explain the following assert?
> 
> 1582    storeDiskdShmGet(SwapDir * sd, int *shm_offset)
> 1583    {
> 1584        char *buf;
> 1585        diskdinfo_t *diskdinfo = sd->fsdata;
> 1586        buf = linklistShift(&diskdinfo->shm.stack);
> 1587        assert(buf);
> 
> 
> What does it tell if it triggers? (which it does for me).
> 
> Am I correct in assuming this is when you run out of shm memory?

Squid allocates one shared memory buffer for each cache_dir.

The size of the buffer is SHMBUFS * SHMBUF_BLKSZ.  Its then
chopped into SHMBUFS chunks, which are put into the
linked list stack.  

If you hit this assertion, it (probably) means there are more
than SHMBUFS I/O's pending.  If you get a core dump, you
can print diskd_stats.

Duane W.


[prev in list] [next in list] [prev in thread] [next in thread]