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

List:       varnish-bugs
Subject:    Re: [Varnish] #1739: overflow on "o ws - Assert error in VFP_Push(),  cache/cache_fetch_proc.c line 
From:       "Varnish" <varnish-bugs () varnish-cache ! org>
Date:       2016-01-27 18:14:06
Message-ID: 058.573e402e2f06cdcb7605684d137e63d2 () varnish-cache ! org
[Download RAW message or body]

#1739: overflow on "o ws - Assert error in VFP_Push(), cache/cache_fetch_proc.c
line 200:
----------------------+-----------------------
 Reporter:  slink     |       Owner:  slink
     Type:  defect    |      Status:  reopened
 Priority:  normal    |   Milestone:
Component:  varnishd  |     Version:  trunk
 Severity:  normal    |  Resolution:
 Keywords:            |
----------------------+-----------------------
Changes (by slink):

 * status:  closed => reopened
 * resolution:  invalid =>


Comment:

 This issue still exists at least for {{{workspace_backend ==
 http_resp_size}}}, reproducible on Solaris with a backend on loopback (may
 be relevant because the default loopback MTU on solaris is 8k):

 {{{V1F_FetchRespHdr}}} calls {{{SES_RxInit}}} with
 {{{cache_param->http_resp_size}}}, so if the actual backend read is large
 enough to use up the reservation on {{{htc->ws == bo->ws}}}, there may be
 no space left on the ws for the VFPs.

 Maybe we should estimate the space required by the VFPs and reduce the
 SES_RxInit allocation by that amount.

-- 
Ticket URL: <https://www.varnish-cache.org/trac/ticket/1739#comment:5>
Varnish <https://varnish-cache.org/>
The Varnish HTTP Accelerator

_______________________________________________
varnish-bugs mailing list
varnish-bugs@varnish-cache.org
https://www.varnish-cache.org/lists/mailman/listinfo/varnish-bugs
[prev in list] [next in list] [prev in thread] [next in thread]