[squid-users] blocking on requests in swap

From: Brian J. Murrell <brian_at_interlinx.bc.ca>
Date: Fri, 28 Sep 2012 09:29:33 -0400

I've upgraded from 3.0.19 to 3.2.1 due to bug 2305. But it seems I
have simply traded one problem for another.

The problem that I am having is that requesting some URLs from squid
leaves the client waiting, forever. For example:

XXX

If I force a cache reload with squidclient's "-r" the URL will fetch
with no problem.

While squid has taken the request and left the client blocked, some
debug shows the following:

2012/09/28 08:57:26.717| client_side_request.cc(1547) doCallouts: Doing clientInterpretRequestHeaders()
2012/09/28 08:57:26.717| client_side_request.cc(138) ~ClientRequestContext: 0x7f4fe0d25fd8 ClientRequestContext destructed
2012/09/28 08:57:26.717| client_side_request.cc(1603) doCallouts: calling processRequest()
2012/09/28 08:57:26.717| clientStreamRead: Calling 1 with cbdata 0x7f4fe0a75420 from node 0x7f4fe0a75458
2012/09/28 08:57:26.717| storeGet: looking up D4679455D49A4795BBAA7B9AFDE29CF7
2012/09/28 08:57:26.717| store_dir.cc(741) get: got in-transit entry: 229039_at_0=0/0/0/2
2012/09/28 08:57:26.717| clientProcessRequest2: default HIT
2012/09/28 08:57:26.717| StoreEntry::lock: key 'D4679455D49A4795BBAA7B9AFDE29CF7' count=3
2012/09/28 08:57:26.717| UFSSwapDir::reference: referencing 0x7f4fe06851e0 0/229039
2012/09/28 08:57:26.717| store_client::copy: D4679455D49A4795BBAA7B9AFDE29CF7, from 0, for length 4096, cb 1, cbdata 0x7f4fe0a74368

Any ideas as to how to resolve this, short of just dumping my entire
cache and starting from fresh?

Cheers,
b.

Received on Fri Sep 28 2012 - 13:29:50 MDT

This archive was generated by hypermail 2.2.0 : Sat Sep 29 2012 - 12:00:05 MDT