couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benoit Chesneau" <>
Subject trying to increase http performance
Date Mon, 21 Apr 2008 00:01:09 GMT
Hi all,

In view of increasing http performances I just tried this little patch
that TCP_NODELAY flag to the socket, which means that even small
amounts of data will be sent immediately. However I have now 7 tests /
18 that fail mostly because some operations are done twice. I get
these errors

basics	failure	5617ms	

   1. Run with debuggerAssertion failed: == 4
   2. Assertion failed: == 6
   3. Assertion failed: == null

conflicts	success	2176ms	
recreate_doc	error	2498ms	

   1. Run with debuggerException raised:
{"error":"conflict","reason":"Update conflict"}

lots_of_docs	success	3115ms	
multiple_rows	success	3310ms	
large_docs	success	2306ms	
utf8	success	3948ms	
attachments	success	1025ms	
content_negotiation	failure	195ms	

   1. Run with debuggerAssertion failed:
xhr.getResponseHeader("Content-Type") == "application/json"

design_docs	failure	4722ms	

   1. Run with debuggerAssertion failed: db.view("test/no_docs") == null
   2. Assertion failed: db.view("test/no_docs") == null

view_collation	success	24147ms	
view_conflicts	success	3367ms	
view_pagination	success	19200ms	
view_sandboxing	success	2879ms	
view_xml	success	1936ms	
replication	error	40525ms	

   1. Run with debuggerAssertion failed:"foo1") == null
   2. Assertion failed: docB._conflicts === undefined
   3. Assertion failed: docA._rev == docB._rev
   4. Assertion failed: docA._rev == docB._rev
   5. Assertion failed: docA._rev == docB._rev
   6. Assertion failed: docA._rev == docB._rev
   7. Assertion failed:"foo1") == null
   8. Exception raised:
[{couch_rep,do_http_request,3},\n {couch_rep,get_missing_revs,2},\n
{couch_rep,maybe_save_docs,4},\n {couch_rep,'-pull_rep/4-fun-0-',5},\n
{couch_btree,stream_kv_node,6},\n {couch_btree,fold,5},\n
{couch_rep,pull_rep,4},\n {couch_rep,replicate,3}]}"}

etags_head	failure	2999ms	

   1. Run with debuggerAssertion failed: etag == xhr.getResponseHeader("etag")

compact	failure	6298ms	

   1. Run with debuggerAssertion failed: deletesize > originalsize
   2. Assertion failed: deletesize > originalsize

I don't know if I'm right about this flag anyway. If someone has some
clue about it :)


- benoƮt

  • Unnamed multipart/mixed (inline, None, 0 bytes)
View raw message