[erlang-questions] [Q] Mnesia is overloaded

Ladislav Lenart lenartlad@REDACTED
Thu Sep 13 17:41:53 CEST 2007


Scott Lystig Fritchie wrote:
>>>>>> "ll" == Ladislav Lenart <lenartlad@REDACTED> writes:
> 
> ll> Mnesia is overloaded: {dump_log, write_threshold}
> 
> Ladislav: This question appears frequently enough on this list that it
> probably ought to become a member of the FAQ.  A search for that error
> message in this list's archives should show several threads over the
> last couple of years.
> 
> Basic summary: it's a sign of a busy system and is mostly harmless.

Hello,

I've read the latest postings about it but I am still a little
bit confused. For example what does "mostly harmless" from above
mean?

Our mnesia database is now running on one node and is overloaded
for ~20 concurrent writes which seems pretty low to me BTW. And
what about ~1000 of them? How long will this report remain harmless?
Or in other words, what is the maximum frequency of these concurrent
bursts that mnesia running on one node can handle?

What is the main cause of the report? Is it a total number of
concurrent writes? Or is it a total size of all data to be written
(from all the writes)?

I am sorry to bother you all with this issue (again), but words
like "often enough" and "mostly harmless" are too vague for me
to be regarded as answers to my questions.

Thanks in advance,

Ladislav Lenart





More information about the erlang-questions mailing list