On Tue, Feb 18, 2014 at 2:28 PM, Jan Lehnardt wrote:
>
> On 18 Feb 2014, at 17:09 , Paul Davis wrote:
>
>> On Tue, Feb 18, 2014 at 9:30 AM, Adam Kocoloski wrote:
>>> On Feb 18, 2014, at 11:38 AM, Jan Lehnardt wrote:
>>>
On 31 Jan 2014, at 20:08 , Jason Smith wrote:
> However ther
On 18 Feb 2014, at 17:09 , Paul Davis wrote:
> On Tue, Feb 18, 2014 at 9:30 AM, Adam Kocoloski wrote:
>> On Feb 18, 2014, at 11:38 AM, Jan Lehnardt wrote:
>>
>>> On 31 Jan 2014, at 20:08 , Jason Smith wrote:
>>>
However there is a pathological situation where you are
updating fast
On Tue, Feb 18, 2014 at 9:30 AM, Adam Kocoloski wrote:
> On Feb 18, 2014, at 11:38 AM, Jan Lehnardt wrote:
>
>> On 31 Jan 2014, at 20:08 , Jason Smith wrote:
>>
>>> However there is a pathological situation where you are
>>> updating faster than the compactor can run, and you will get an infinit
On Feb 18, 2014, at 11:38 AM, Jan Lehnardt wrote:
> On 31 Jan 2014, at 20:08 , Jason Smith wrote:
>
>> However there is a pathological situation where you are
>> updating faster than the compactor can run, and you will get an infinite
>> loop (plus very heavy i/o and filesystem waste as the com
On 31 Jan 2014, at 20:08 , Jason Smith wrote:
> Anyway I think the broader point is, compaction is for compacting databases
> (removing old document revisions), and replication is for making a copy of
> a database (or subset). If compaction is causing downtime then that is a
> different bug to t
Typo'ed:
"so it's unbalancing as much"
should read:
"so it's *not* unbalancing as much"
On Mon, Feb 17, 2014 at 12:49 PM, Robert Samuel Newson
wrote:
> Replication will not rebalance the tree, no. It's just adding to the end (and
> unbalancing the tree).
>
> The updates are happening in batch
Replication will not rebalance the tree, no. It’s just adding to the end (and
unbalancing the tree).
The updates are happening in batches, though, so it’s unbalancing as much as
the original individual updates did.
B.
On 17 Feb 2014, at 16:16, Boaz Citrin wrote:
> Did some more testing. seem
Did some more testing. seems like indeed compaction is faster than replication.
One thing I observe though is that replication doesn't result the same
as compaction;
While it only copies the leaves, I suspect it doesn't produce a
balanced tree, so subsequent compaction is needed anyway (and indeed
On Jan 31, 2014, at 3:43 PM, Jens Alfke wrote:
> On Jan 31, 2014, at 12:07 PM, Boaz Citrin wrote:
>
>> But if replication only copies the leaf then it makes sense that it is
>> fatser, at least on the same machine. Instead of balancing a tree it just
>> copies a single revision.
>
> Um, no. Th
On Jan 31, 2014, at 12:07 PM, Boaz Citrin wrote:
> But if replication only copies the leaf then it makes sense that it is
> fatser, at least on the same machine. Instead of balancing a tree it just
> copies a single revision.
Um, no. The copied revision has to be inserted into the tree on the t
Boaz, I encourage you to benchmark both techniques. I think you will
discover that compaction is faster. However if not, then you will at least
have peace of mind that your approach is the best.
On Sat, Feb 1, 2014 at 3:07 AM, Boaz Citrin wrote:
> But if replication only copies the leaf then it
On Sat, Feb 1, 2014 at 2:27 AM, Boaz Citrin wrote:
> @Jason - the purpose is that compaction will not catch up with heavy
> write rate and will affect performance.
>
If compaction will not catch up with your writes, then you do not have
sufficient hardware resources. You are operating in a very
But if replication only copies the leaf then it makes sense that it is
fatser, at least on the same machine. Instead of balancing a tree it just
copies a single revision.
Also I undestand from Jason that compaction tries to catch up, unlike a non
continuos replication.
So it seems that if I don't n
On Jan 31, 2014, at 11:27 AM, Boaz Citrin wrote:
> @Jason - the purpose is that compaction will not catch up with heavy
> write rate and will affect performance.
The same will be true of replication.
> I wonder if I can omit the compaction if replication gives the same result.
> Will replicati
@Jason - the purpose is that compaction will not catch up with heavy
write rate and will affect performance.
So I replicate/copy to a side database, compact it, then replicate the
new changes from the original and switchover.
I wonder if I can omit the compaction if replication gives the same resul
Anyway I think the broader point is, compaction is for compacting databases
(removing old document revisions), and replication is for making a copy of
a database (or subset). If compaction is causing downtime then that is a
different bug to talk about, but it should be totally transparent.
Jens (i
On Jan 31, 2014, at 9:46 AM, Mark Hahn wrote:
> It wouldn't matter if it did. Within the same server linux short-circuits
> http to make it the same as unix sockets, i.e. very little overhead.
I think you mean it short-circuits TCP :)
There's extra work involved in HTTP generation & parsing no
> replication within the same CouchDB server does not use HTTP
It wouldn't matter if it did. Within the same server linux short-circuits
http to make it the same as unix sockets, i.e. very little overhead.
On Fri, Jan 31, 2014 at 9:43 AM, Jason Smith wrote:
> On Sat, Feb 1, 2014 at 12:30 AM,
On Sat, Feb 1, 2014 at 12:30 AM, Jens Alfke wrote:
>
> On Jan 31, 2014, at 8:39 AM, Boaz Citrin wrote:
>
> > Instead we could replicate the database into a new instance and
> > switchover, as replication is much faster.
>
> That seems sort of strange to me — I would expect replication to an empt
On Jan 31, 2014, at 8:39 AM, Boaz Citrin wrote:
> Instead we could replicate the database into a new instance and
> switchover, as replication is much faster.
That seems sort of strange to me — I would expect replication to an empty db to
be slower, as it's doing basically the same work as com
Hello,
We need to periodically need to compact our database.
The problem is that this takes many hours and we cannot stay offline
for so long.
Instead we could replicate the database into a new instance and
switchover, as replication is much faster.
The question however is if both produces the sam
21 matches
Mail list logo