2009/5/19 Mark Nottingham :
> I'm going to push back on that; the administrator doesn't really have any
> need to get a core when, for example, append_domain doesn't start with .'.
>
> Squid.conf is bloated as it is; if there are cases where a core could be
> conceivably useful, they should be conv
Can you craft a small C program to replicate the behaviour?
adrian
2009/5/24 Guido Serassio :
> Hi,
>
> One user has reported a very strange problem using cache_peer directive on
> 2.7 STABLE6 running on Windows:
>
> When using the following config:
>
> cache_peer 192.168.0.63 parent 3329 0
ded to getservbyname() under Windows. :)
adrian
2009/5/24 Guido Serassio :
> Hi,
>
> At 04.38 24/05/2009, Adrian Chadd wrote:
>>
>> Can you craft a small C program to replicate the behaviour?
>
> Sure, I wrote the following test program:
>
> #include
> #in
strtoul(). But if you want to verify the -whole- thing is numeric,
just write a bit of C which does this:
int isNumeric(const char *str)
{
}
2009/5/25 Amos Jeffries :
> Guido Serassio wrote:
>>
>> Hi,
>>
>> At 16.17 24/05/2009, Adrian Chadd wrote:
>>>
>&
int
isUnsignedNumeric(const char *str)
{
for (; *str; str++) {
if (! isdigit(*str))
return -1;
}
return 1;
}
2009/5/25 Adrian Chadd :
> strtoul(). But if you want to verify the -whole- thing is numeric,
> just write a bit of C which does this:
>
> i
Sorry!
Adrian
2009/5/25 Kinkie :
> On Mon, May 25, 2009 at 2:21 PM, Adrian Chadd wrote:
>> int
>> isUnsignedNumeric(const char *str)
>> {
>> for (; *str; str++) {
>> if (! isdigit(*str))
>> return -1;
>> }
>> return 1;
I'm working on a couple of paid squid + active directory deployments
and they're both seeing the occasional NTLM auth popup happening.
The workaround is pretty simple - just enable the IP auth cache. This
however doesn't solve the fundamental problem(s), whatever they are.
The symptom is logs lik
G'day guys,
I've fixed a bug in Lusca which was introduced with Benno's method_t
stuff. The specific bug is revalidation replies 'hanging' until the
upstream socket closes, forcing an end of message to occur.
The history and patch are here:
http://code.google.com/p/lusca-cache/source/detail?r=141
2009/7/15 Ian Hickson :
> On Tue, 14 Jul 2009, Alex Rousskov wrote:
>>
>> WebSocket made the handshake bytes look like something Squid thinks it
>> understands. That is the whole point of the argument. You are sending an
>> HTTP-looking message that is not really an HTTP message. I think this is
>>
2009/7/15 Amos Jeffries :
> a) Getting a dedicated WebSocket port assigned.
> * You and the client needing it have an argument to get that port opened
> through the firewall.
> * Squid and other proxies can be altered to allow CONNECT through to safe
> defined ports (80 is not one). Or to do t
2009/7/16 Ian Hickson :
> We actually used to do that, but we got requests to make it more
> compatible with the HTTP Upgrade mechanism so that people could add the
> support to their HTTP servers instead of having to put code in front of
> their servers.
Right. So why not extend the spec a little
2009/7/16 Ian Hickson :
>> Right down to the HTTP/1.1 reserved protocol label (do get that changed
>> please).
>
> If we're faking HTTP, then it has to look like HTTP.
The message here is "don't fake HTTP". "Speak HTTP over port 80".
> I'm getting very mixed messages here.
>
> Is there a reliabl
NOte that winbind has a hard coded limit that is by default very low.
Opening 2n ntlm_auth helpers may make things blow up in horrible ways.
Adrian
2009/7/16 Robert Collins :
> On Thu, 2009-07-16 at 14:08 +1200, Amos Jeffries wrote:
>>
>> Both reconfigure and helper recovery use startHelpers()
2009/7/17 Ian Hickson :
>> That way you are still speaking HTTP right until the "protocol change"
>> occurs, so any and all HTTP compatible changes in the path(s) will
>> occur.
>
> As mentioned earlier, we need the handshake to be very precisely defined
> because otherwise people could trick unsu
G'day,
I've fixed a potentially risky situation in Lusca relating to the
initialisation of the storeIOState cbdata type. Each storedir has a
different idea of how the allocation should be free()'ed.
The relevant commit in Lusca is r14208 -
http://code.google.com/p/lusca-cache/source/detail?r=1420
2009/7/20 Henrik Nordstrom :
>> I've fixed a potentially risky situation in Lusca relating to the
>> initialisation of the storeIOState cbdata type. Each storedir has a
>> different idea of how the allocation should be free()'ed.
>
> Risky in what sense?
Ah. I just re-re-re-read the code again an
G'day,
I just noticed in src/HttpReply.c that the vary expire option
(Config.onoff.vary_ignore_expire) is checked if the reply has HDR_VARY
set but it does not check if HDR_X_ACCELERATOR_VARY is set.
Everywhere else in the code checks them both consistently and
assembles "Vary" header contents c
2009/8/16 Henrik Nordstrom :
> sön 2009-08-16 klockan 10:23 +1000 skrev Robert Collins:
>
>> If the noise is too disturbing to folk we can investigate these... I
>> wouldn't want anyone to leave the list because of these reports.
>
> I would expect the number of reports to decline significantly as
G'day. This question is aimed mostly at Henrik, who I recall replying
to a similar question years ago but without explaining why.
Why does Squid-2 return HTTP_PROXY_AUTHENTICATION_REQUIRED on a denied ACL?
The particular bit in src/client_side.c:
int require_auth = (answer == ACCESS_REQ_PROXY_AU
Guys,
Please look at what other multi-CPU network applications do, how they
work and don't work well, before continuing this kind of discussion.
Everything that has been discussed has already been done to death
elsewhere. Please don't re-invent the wheel, badly.
Adrian
2009/9/15 Robert Collin
2009/9/15 Sachin Malave :
> On Tue, Sep 15, 2009 at 1:18 AM, Adrian Chadd wrote:
>> Guys,
>>
>> Please look at what other multi-CPU network applications do, how they
>> work and don't work well, before continuing this kind of discussion.
>>
>> Everyth
But in that case, ACCESS_REQ_PROXY_AUTH would be returned rather than
ACCESS_DENIED..
Adrian
2009/9/15 Robert Collins :
> On Tue, 2009-09-15 at 15:22 +1000, Adrian Chadd wrote:
>> G'day. This question is aimed mostly at Henrik, who I recall replying
>> to a similar qu
e
notification between threads.
You can then push some "stuff" into these worker threads as an
experiment and see exactly what the issues are.
Building worker threads into Squid is easy. Making them do anything?
Not so easy :)
Adrian
2009/9/15 Sachin Malave :
> On Tue, Sep 15, 2009 at
Hi,
After yet another long break I'm kind of back.
Expect bits and pieces of things to start coming out..
:)
Adrian
Hiya,
I'd like to remove the fd_set stuff from delay pools.
My first hack was to turn the fd_set's into arrays.
That looked fine. :-)
Do people have an issue with this? The downside is the
byte-per-fd wasted but considering how much RAM squid normally
uses I don't see it as an issue.
What do pe
On Thu, Feb 27, 2003, Robert Collins wrote:
> > What do people think?
>
> I'm working on a more generic solution at the moment, consolidating
> deferred reads and delay pools etc..
>
> Rough notes in bugzilla, or pop onto IRC and we can discuss...
Can I have a look at what you've done?
On Thu, Feb 27, 2003, Robert Collins wrote:
> Sure. I'm pushing the arch mirror now. It's in the epoll branch
> ([EMAIL PROTECTED]/squid--epoll--3.0).
>
[snip]
It sounds good. Now, what about squid-2.5? :)
Adrian
On Thu, Feb 27, 2003, Robert Collins wrote:
> >
> > It sounds good. Now, what about squid-2.5? :)
>
> 2.5? whassat?
What people are still running in release? :)
adrian
On Thu, Feb 27, 2003, Henrik Nordstrom wrote:
> What is the reason to this?
>
> I guess it is your bugzilla entry..
Issues with FD_SET and this glibc under Linux.
This fixed crashes w/ delay pools that have been happening for months.
(As far as I know, I'll run it for a few more days..)
Adria
On Fri, Feb 28, 2003, Henrik Nordstrom wrote:
> Adrian: What was the exact number of filedescriptors you configured
> Squid for using?
2048.
> I am a bit reluctant about doing this change in Squid-2.5 as to me
> this large (but non-intrusive) change does not seem to be needed if
> the FD_SETSI
Hi,
Here's another one. I've just hit an issue where a pair of squids are
in a magical redundant setup (if the master fails, the backup takes
over the IP).
The measure of "failed" is "http port isn't bound".
Now, the -F option to squid makes it rebuild faster BUT it still binds to
the socket. I
On Mon, Mar 03, 2003, Henrik Nordstrom wrote:
> m?n 2003-03-03 klockan 11.59 skrev Robert Collins:
>
> > It
> > * removes deferred reads across the board
> > * converts nearly all unready reads to the new DeferredRead framework.
> > * leaves ssl reading 1 byte per IO cycle at minimum.
> > * Includ
Yo,
source='ufs/store_dir_ufs.cc' object='ufs/store_dir_ufs.o' libtool=no
depfile='.deps/ufs/store_dir_ufs.Po' tmpdepfile='.deps/ufs/store_dir_ufs.TPo'
depmode=gcc /usr/local/bin/bash ../../cfgaux/depcomp g++ -DHAVE_CONFIG_H -I. -I.
-I../../include -I. -I../../include -I../../include -I..
yo,
building with no delay pools doesn't work:
[EMAIL PROTECTED]:~/work/squid/squid-cvs/squid3/src> make store.o
source='store.cc' object='store.o' libtool=no depfile='.deps/store.Po'
tmpdepfile='.deps/store.TPo' depmode=gcc /usr/local/bin/bash ../cfgaux/depcomp g++
-DHAVE_CONFIG_H -DDEFAUL
Robert knows about this, but just FYI
At the moment squid-3 is leakiing short strings all over the floor.
I'm currently polygraphing a squid + kqueue w/ a null store fs.
I'll check out the storefs stuff once I'm happy with the non-storefs
run. :)
Adrian
Hiya,
just send an email to [EMAIL PROTECTED] .
I'll approve you. :)
Adrian
On Wed, Mar 05, 2003, [EMAIL PROTECTED] wrote:
> Hi everybody,
> my name is Michele De Martin and I'm working as a network administrator:
> routers, switches but also DNSs and proxies services.
> I've some years of ex
hiya,
I'm in a position where a few clients are complaining about the mod_gzip
Vary: ignore and I'd like to return valid replies to them.
I spoke to Robert on IRC about this a little. I thought about hacking up
the Vary: header wherever I discovered there was a mod_gzip reply w/ no
vary-informat
ok,
squid-3 doesn't seem to be leaking memory everywhere now, but:
2003/03/06 21:59:54| assertion failed: cbdata.cc:318: "c->locks < 65535"
I'm going to run it again overnight, see if it triggers. If so,
I'm going to have to hack it up to not exit cleanly - I didn't get a stack trace. :)
Ad
Program received signal SIGABRT, Aborted.
0x281edb58 in kill () from /usr/lib/libc.so.4
(gdb) bt
#0 0x281edb58 in kill () from /usr/lib/libc.so.4
#1 0x2822f10a in abort () from /usr/lib/libc.so.4
#2 0x8085307 in xassert (msg=0x80fc925 "c->locks < 65535", file=0x80fc805
"cbdata.cc", line=318)
On Thu, Mar 06, 2003, Henrik Nordstrom wrote:
> > (gdb)
> >
> > hm. I don't remember ever doing that. :) I'll take a squizz in a tick.
>
>
> cbdataReferenceDone missing from ConnStateData::~ConnStateData I think..
>
I'll leave it for robert to look at tomorrow morning (thismorning. :)
That w
hiya,
Squid-3 seems to have settled down quite a bit now. Its been doing this 125req/sec
for a couple of hours now with no visible memory leaks. We'll have to stress out
some of the ACL stuff to check for leaks there and I'll turn on ufs and diskd
to see how they fare.
So far, so good. :-)
Robe
On Fri, Mar 07, 2003, Robert Collins wrote:
> > I'm going to run it again overnight, see if it triggers. If so,
> > I'm going to have to hack it up to not exit cleanly - I didn't get a stack trace.
> > :)
>
> Thanks, fixed.
Nope. Still triggered.
Although, it _did_ take a lot longer to trigger
On Thu, Mar 06, 2003, Adrian Chadd wrote:
> Nope. Still triggered.
>
> Although, it _did_ take a lot longer to trigger this time.
>
> Let me start it up again with full cb debugging, try to get a history on the thing.
.. which I think I've squished. Let me run the polygraph magic.
Adrian
On Thu, Mar 06, 2003, Adrian Chadd wrote:
> On Thu, Mar 06, 2003, Adrian Chadd wrote:
> > Nope. Still triggered.
> >
> > Although, it _did_ take a lot longer to trigger this time.
> >
> > Let me start it up again with full cb debugging, try to get a history on
hi,
squid3 installs diskd in libexec/diskd/diskd/ but diskd_program
wants libexed/diskd.
Can someone take a look? :)
Adrain
yo,
here's a diskd bug I'm seeing:
2003/03/07 14:28:56| could not parse headers from on disk structure!
.. lots and lots of em.
I'll try with UFS in a minute.
Adrian
On Thu, Mar 06, 2003, Adrian Chadd wrote:
>
> yo,
>
> here's a diskd bug I'm seeing:
>
> 2003/03/07 14:28:56| could not parse headers from on disk structure!
>
> .. lots and lots of em.
>
> I'll try with UFS in a minute.
>
So far 75req/sec on UFS, no message. Hm. Diskd evilness?
Adrian
On Sun, Mar 09, 2003, Robert Collins wrote:
> > This happens on all Windows build environments: Cygwin, MinGW and MS Visual
> > Studio.
>
> Dang. It's just *so hard* with MS polluting the namescape all the time.
>
> If we don't use the windows ACL type, just do this:
>
Heh. Prefix everything
On Mon, Mar 10, 2003, David Nicklay wrote:
> Brian and I have been trying to puzzle out a problem we are having
> related to back end origin server connections initiated by squid. We
> have squid (2.5.stable1) set up in a reverse proxy configuration
> pointing at a group of origin servers which
On Wed, Mar 12, 2003, Henrik Nordstrom wrote:
> But why is it that we keep running into automake bugs all the time?
> The use of automake is supposed to cause less grief, not more..
Heh. Bleeding edge features..
btw, check this out:
(gdb) run -ND
Starting program: /home/adrian/work/squid/squid
On Wed, Mar 12, 2003, Henrik Nordstrom wrote:
> Hmm.. maybe the comm abort logics is not as fool proof as I thought.
>
> Here did I put that patch for adding cbdata fences to comm.. probably
> lost.
>
> But on the other hand this looks like it could be a httpState, but
> slightly stomped by some
Hi,
A few accounting-style parsing scripts here at uni are breaking because
there's a few ad-programs which try authenticating with a blank username.
I poked Robert about it, he thinks a blank username should just be logged
as -. Obviously the auth attempt fails.
(I don't knwo whether its a bug
On Mon, Jul 28, 2003, Henrik Nordstrom wrote:
>
> Better to make accessLogFormatName return NULL on blank names I think.
> The logics in all log functions assumes NULL == unknown, and
> addressing this in accessLogFormatName would cover all bases in one
> clean go.
How about this:
accessLogF
Hi,
I'm doing some testing w/ COSS under linux and I'm suffering issues
w/ POSIX AIO. My main problem is zombie/defunct processes, which
I can only think is due to bugs in the posix AIO / thread implementation.
adrian1072 0.0 0.0 00 pts/1ZAug05 0:00 [squid ]
adrian107
On Tue, Aug 05, 2003, Henrik Nordstrom wrote:
> On Tuesday 05 August 2003 18.05, Adrian Chadd wrote:
> > Hi,
> >
> > I'm doing some testing w/ COSS under linux and I'm suffering issues
> > w/ POSIX AIO. My main problem is zombie/defunct processes, which
>
On Mon, Aug 11, 2003, Robert Collins wrote:
> >Bug #638 assertion failure if proxy_auth used wrongly in
> > delay_access
>
> Hmm, little motivation here - proxy_auth is fully supported in 3.0.
Yes, but I'm sure we're going to find people running squid-2.5 for
the next year or two. C.f. squi
On Thu, Aug 14, 2003, Robert Collins wrote:
>
> It hasn't changed that much!
Yeah, I'm looking at it a little more. heh.
> Seriously, the stack trace shows httpSendRequest as the active culprit,
> and in frame 10 you should be able to get the address of the callback
> listed for the previous wr
Any ideas? I forgot how to fix this:
(gdb) print fd
$1 = 28
(gdb) print fd_table[28].wstate
cannot subscript something of type `'
(gdb) print fd_table[28].write_method
cannot subscript something of type `'
(gdb)
...?
adrian
On Thu, Aug 14, 2003, Robert Collins wrote:
> On Thu, 2003-08-14 at 23:14, Adrian Chadd wrote:
>
> It's a duplicate write - a write had been scheduled, not triggered, and
> now a new write is scheduled.
Yeah, its the old school IO code. I mean, any suggestions on whats
causing
On Wed, Aug 13, 2003, Adrian Chadd wrote:
> On Thu, Aug 14, 2003, Henrik Nordstrom wrote:
>
> > It would be very undesireable from a release management perspective if
> > COSS in Squid-2.5 is newer than COSS in Squid-3.0, and neither me or
> > Robert will be very happy i
On Thu, Aug 14, 2003, Henrik Nordstrom wrote:
> It would be very undesireable from a release management perspective if
> COSS in Squid-2.5 is newer than COSS in Squid-3.0, and neither me or
> Robert will be very happy if this happens.
Let me try to make it work first. heh.
Adrian
A bit late, yeah, I've been a little busy until recently.
On Mon, Apr 28, 2003, Alex Kinch wrote:
> I've not really got any experience of code hacking, think that's best left
> to the experts on here (Hi Adrian - remember Amsterdam?!). However, I can
> give plenty of feedback on the ESI function
On Thu, Aug 14, 2003, Robert Collins wrote:
> Let me add:
> It will be a -lot- easier for you to:
> Get 3.0 stable on those uni caches.
> Get kqueue / epoll stable on those uni caches.
> Then get COSS stable on 3.0,
> than to:
> get COSS stable.
> get 3.0 stable
> get coss with 3.0 stable
> get c
On Thu, Aug 14, 2003, Henrik Nordstrom wrote:
> OK for me. COSS is an experimental feature and not under code freeze
> even for Squid-2.5. Just remember to update the patches page after
> commit.
I'll do that once I've finished my rebuild code.
> If you need to touch any files outside of src/f
On Wed, Aug 06, 2003, Henrik Nordstrom wrote:
> You might be able reap them by doing a waitpid(-1, NULL, WNOHANG) in
> the main loop somewhere..
>
Tried, no love. I think there's dragons.
> In any case, file a bug report with your vendor.
Or upgrade to debian-unstable w/
Linux mierda 2.6.0-
hi all,
I've recently been working a little more on COSS. Thanks to Duane and his
blocksize work its actually reasonably useful now. I'm having issues
tuning it (since it does waste quite a bit of space, eep!) _but_ the
performance increases make up for it.
I've got some initial patches which in
squid-3, 2gig ufs, 30req/sec datacomm-1 run. Left it for about an hour,
then ctrl-C'ed the polygraph client/server. This happened:
(gdb) bt
#0 0x401b9a41 in kill () from /lib/libc.so.6
#1 0x4005224b in pthread_kill () from /lib/libpthread.so.0
#2 0x40052521 in raise () from /lib/libpthread.so.
On Thu, Aug 14, 2003, Robert Collins wrote:
> There are multiple stripes per coss 'dir' right? So, perhaps you write a
> tag records to the swaplog for the coss dir after the normal records,
> listing the status of the stripes?
Thats what I've been saying, but there's a few gotchas which I've co
On Thu, Aug 14, 2003, Robert Collins wrote:
> On Thu, 2003-08-14 at 16:31, Adrian Chadd wrote:
>
> > There's no way to do a dirty rebuild unless:
>
> Easy answer: a dirty coss rebuild is simply to purge the stripes that
> are dirty.
Well, at the moment, there'
On Fri, Aug 15, 2003, Robert Collins wrote:
>
> you built with -g ?
-g -ggdb. Unless, of course, the damned linker didn't link it
with -g -ggdb...
adrian
On Fri, Aug 15, 2003, Adrian Chadd wrote:
> On Fri, Aug 15, 2003, Robert Collins wrote:
> >
> > you built with -g ?
>
> -g -ggdb. Unless, of course, the damned linker didn't link it
> with -g -ggdb...
>
...
(gdb) frame 7
#7 0x0808483e in comm_old_
On Fri, Aug 15, 2003, Robert Collins wrote:
> [EMAIL PROTECTED]:~/source/squid/squid--HEAD--3.0$ grep fde_t src/*.cc
>
> I think you'll have more luck with
> print (fde *)fd_table[fd]
Yeah, I'm stupid. It worked, but having to have casts...
Adrain
On Thu, Aug 21, 2003, Robert Collins wrote:
> On Thu, 2003-08-21 at 05:45, Duane Wessels wrote:
>
> > FYI, my coss changes have not been commited to squid-3 yet because
> >
> >- I was going to wait until the time when it is okay to
> > commit non-bugs to the tree.
>
> And on this note -
On Wed, Aug 27, 2003, Kinkie wrote:
> In fact, it's a common trend in current RDBMs: they're starting to favour
> storage areas on fileystems rather than on raw devices. Linux has better
> support for filesystem-based raw operations (O_DIRECT) than to raw devices
> proper.
>
> BTW: O_DIRECT would
On Wed, Aug 27, 2003, Henrik Nordstrom wrote:
> On Wednesday 27 August 2003 09.51, Kinkie wrote:
>
> > BTW: O_DIRECT would be a good approach to our FS I/O work but:
> > - we need to be able to keep an object hot after reading it - the
> > OS won't do it for us anymore
>
> This is already on the
On Wed, Aug 27, 2003, Duane Wessels wrote:
> > Duane, would you be willing to forward port your squid-2.5 COSS work
> > to squid-3? I'm currently evaluating Squid-3 for a cache here and
> > I'd like to test both epoll and COSS out.
>
> Done. I've committed the code to squid-3. It compiles, but I
Here's the current stats from a squid-3 test. I'm not terribly impressed.
Throughput wise, AUFS is about the same as squid-2.5. No surprises there.
I'll be interested in seeing how squid-3+coss and squid-3+epoll performs.
But, squid seems to be taking an _awfully_ large amount of RAM: ie,
adria
On Sat, Aug 30, 2003, Robert Collins wrote:
> The only step missing to be able to use the IO modules from anywhere is
> automatic registration to be 'synced' at shutdown, and 'callback()'
> checked on each io loop.
If you want to use the async io code rom anywhere, say, logfile writing,
you're _g
On Sat, Aug 30, 2003, Serassio Guido wrote:
> >changes that can be improved upon, etc.
>
> This is a very interesting thing:
>
> For example it means that a Windows native awin32 fs based on aufs is no
> more needed: all the magic can will be do from a Win32DiskThreads.
Or, once COSS has been
On Sun, Aug 31, 2003, Robert Collins wrote:
> > If you want to use the async io code rom anywhere, say, logfile writing,
> > you're _going_ to have to do write combining or your write queue is going
> > to quickly blow out to a stupid size. The store IO interface really assures
> > us there's only
On Thu, Sep 04, 2003, Henrik Nordstrom wrote:
> I am currently working in a development works in epoll I/O on squid for satisfying
> the requests of 2000 in IA64.I have passed only 300 req/sec in IA64.My squid is
> using more than 1.9 GB outof 2 GB.I have tried a lot in the squid to overcome the
On Sat, Sep 06, 2003, Pawel Worach wrote:
> Looks like the swap path gets trashed along the way,
> this was triggered by squid -z with debug level above 0.
Heh. i just committed a similar patch to deal with (null) showing up.
Hm,
I'm now successfully running squid-3 on an older machine running
a 2.6 kernel. I'm putting my memory usage issues down to
"random mismatched libraries, debian kthx." Oh well.
I'll give the aufs and coss code a decent run and then I'll
start improving it.
Adrian
On Wed, Sep 10, 2003, Henrik Nordstrom wrote:
> It is my opinion that there is many better ways of optimizing proxy
> class I/O without resorting to kernel level "splicing". For example
> it would be very interesting to see a I/O mechanism not involving
> copying of data on each read/write comb
hi,
I'm running a snapshot from a couple days ago and I'm getting:
2003/09/21 19:24:02| could not parse headers from on disk structure!
2003/09/21 19:24:02| clientReplyContext::sendMoreData: Unable to parse reply headers
within a single HTTP_REQBUF_SZ length buffer
2003/09/21 19:24:02| could no
On Sun, Sep 21, 2003, Robert Collins wrote:
> > .. any ideas?
>
> possibly related to the stmem "0" assert, -or- you have a bunch of
> corrupt swapfiles.
Well, its triggering that assertion. Hm, is it fixed in the latest
CVS codebase?
> You might check via ufsdump the data for those files.
I wo
On Mon, Sep 22, 2003, Robert Collins wrote:
> > during swapin..
> >
> > Maybe it is the same problem?
>
> Possibly as a triggering cause, yes.
>
> Adrian,
> I've committed an extension to the assert, so that triggering it will
> dump the available byte ranges from the mem_hdr.
>
When I
hi all,
Here's a fix for a memory leak I found during some
polygraph stress testing:
Index: store_swapout.cc
===
RCS file: /server/cvs-server/squid/squid3/src/store_swapout.cc,v
retrieving revision 1.97
diff -u -r1.97 store_swapout.
On Thu, Feb 12, 2004, Robert Collins wrote:
>
> please commit.
Done.
Next leak..
Adrian
Hi all,
I've got myself a little memory leak. We're leaking the fwdState
mempool allocations. A leaked allocation looks like this:
2004/02/11 16:00:12.976| cbdataAlloc: 0x84e2010 forward.cc:914
2004/02/11 16:00:12.976| cbdataLock: 0x84e2010=1 peer_select.cc:169
2004/02/11 16:00:12.977| cbdataRef
Right. Here's a little exercise into .. well, a few examples of badness
in C/C++ code.
We're seeing a memory leak in the fwdState cbdata allocation. After
a bit of tracing I discovered this routine:
void
ConnectStateData::callCallback(comm_err_t status, int xerrno)
{
debug(5, 3) ("commCon
Hi,
Here's a little patch which does three things:
* breaks out the if (cbdataReferenceValid(data)) check
which is using CallBack.data into a seperate method;
* deletes the extra reference/dereference which the CallBack
code is meant to be doing for us; and
* uses the two above in the offend
On Sun, Feb 15, 2004, Adrian Chadd wrote:
>
> Hi,
>
> Here's a little patch which does three things:
.. needs to do four:
Index: comm.cc
===
RCS file: /server/cvs-server/squid/squid3/src/comm.cc,v
retrieving re
Another crash, this time, reproducable:
2004/02/16 16:27:31| assertion failed: comm.cc:2558: "fdc_table[fd].active &&
!fdc_table[fd].half_closed"
Program received signal SIGABRT, Aborted.
0x40186da9 in raise () from /lib/tls/libc.so.6
(gdb)
(gdb) bt
#0 0x40186da9 in raise () from /lib/tls/libc
On Mon, Feb 16, 2004, Adrian Chadd wrote:
>
> Another crash, this time, reproducable:
>
> 2004/02/16 16:27:31| assertion failed: comm.cc:2558: "fdc_table[fd].active &&
> !fdc_table[fd].half_closed"
>
> Program received signal SIGABRT, Aborted.
> 0x40
Hi,
I'd like to commit this to the squid-3 tree.
Can those who are seeing some memory leaks please update and
apply this patch?
Adrian
Index: comm.cc
===
RCS file: /server/cvs-server/squid/squid3/src/comm.cc,v
retrieving revisio
On Mon, Feb 16, 2004, Adrian Chadd wrote:
> On Mon, Feb 16, 2004, Adrian Chadd wrote:
> >
> > Another crash, this time, reproducable:
> >
> > 2004/02/16 16:27:31| assertion failed: comm.cc:2558: "fdc_table[fd].active &&
> > !fdc_table[fd].half_close
On Mon, Feb 16, 2004, Adrian Chadd wrote:
> On Mon, Feb 16, 2004, Adrian Chadd wrote:
> > On Mon, Feb 16, 2004, Adrian Chadd wrote:
> > >
> > > Another crash, this time, reproducable:
> > >
> > > 2004/02/16 16:27:31| assertion failed: comm.cc:2558:
Hi,
I'm playing with the epoll code in the latest libc+linux 2.6 kernel
and I've found that we don't need the -lepoll anymore. Its
definitely integrated into the newest libc.
Should we remove the -lepoll? Make configure test for its presence
if --enable-epoll is given?
Adrian
HI,
I've fixed, hopefully, a large chunk of the memory leaks people
have been seeing in squid-3-cvs. I suggest upgrading any installs
you have to the latest and give it a run.
I'm currently stressing out squid-3-cvs with epoll and diskd.
Adrian
501 - 600 of 1104 matches
Mail list logo
