[squid-users] Second COSS swap is unused?
Hello! I have following setup on my squid: cache_dir aufs /var/cache/squid 31000 32 256 cache_dir coss /var/cache/squid/COSS.img 2048 max-size=131072 cache_dir aufs /var/cache/squid2 31000 32 256 cache_dir coss /var/cache/squid2/COSS.img 2048 max-size=1048575 But second COSS seems to be unused: Store Directory #1 (coss): /var/cache/squid/COSS.img FS Block Size 512 Bytes Maximum Size: 2097152 KB Current Size: 1908943 KB Percent Used: 91.03% Current load metric: 3 / 1000 Number of object collisions: 2 Flags: SELECTED Pending Relocations: 0 Stripe: 1032, lockcount: 0, numobjects 62, flags: NOTFULL,NOTWRITING,NOTWRITTEN,DISK Stripe: 0, lockcount: 0, numobjects 135, flags: NOTFULL,NOTWRITING,NOTWRITTEN,MEMONLY Stripe: 3, lockcount: 1, numobjects 288, flags: FULL,NOTWRITING,NOTWRITTEN,MEMONLY Pending Relocations: Store Directory #3 (coss): /var/cache/squid2/COSS.img FS Block Size 512 Bytes Maximum Size: 2097152 KB Current Size: 0 KB Percent Used: 0.00% Current load metric: 0 / 1000 Number of object collisions: 0 Flags: Pending Relocations: 0 Stripe: 0, lockcount: 0, numobjects 0, flags: NOTFULL,NOTWRITING,NOTWRITTEN,MEMONLY Stripe: 0, lockcount: 0, numobjects 0, flags: NOTFULL,NOTWRITING,NOTWRITTEN,DISK Pending Relocations: What I am doing wrong? Maybe I should remove two COSS swaps and make one? -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Second COSS swap is unused?
Adrian Chadd wrote: On Thu, Sep 21, 2006, Micha? Margula wrote: Hello! I have following setup on my squid: cache_dir aufs /var/cache/squid 31000 32 256 cache_dir coss /var/cache/squid/COSS.img 2048 max-size=131072 cache_dir aufs /var/cache/squid2 31000 32 256 cache_dir coss /var/cache/squid2/COSS.img 2048 max-size=1048575 Re-order the disks so the two COSS stripes are first. Thank you. That helped. I had to restart squid, squid -k reconfigure did not change order. Now I will sit and watch how those files fill up, because I still don't know how big to make them. But even with only one coss file used, Load Average dropped below 1.0 (average for last 8 hours is 0.3, usually average was about 4!). Thanks! -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
[squid-users] IOWAIT and second disk
Hello! I have a aufs on /dev/sdc. Iowait takes about 40% of CPU. If I add another disk, will it make that load lower? Is it good idea to use two disks with aufs (diskd is unstable unfortunately)? -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] IOWAIT and second disk
Adrian Chadd wrote: On Tue, Sep 19, 2006, Micha? Margula wrote: Hello! I have a aufs on /dev/sdc. Iowait takes about 40% of CPU. If I add another disk, will it make that load lower? Is it good idea to use two disks with aufs (diskd is unstable unfortunately)? Yes, AUFS is fine. Take a look at http://www.squid-cache.org/~adrian/coss/ - note the dramatic drop in IOWAIT when using COSS. It also depends on the hardware you're using. What kind of disks are they? Which OS? Which controller? Linux 2.6.15.1, /dev/sdc with cache is Seagate ST336754LC (Cheetach) - 36 GB, 15k RPMs, and logging is done on /dev/md0 which is RAID1 consisting of two disks like the one above. But I don't think RAID1 causes any trouble, iostat shows majority of IO operations on /dev/sdc where cache resides. Hardware is 2 x Xeon 3.6, 2GB RAM, SCSI storage controller: LSI Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI on Intel E7520. Now in topic of COSS, what is better solution (/dev/sdc is 36 GB and /dev/sdd also will be 36 GB, same disk model): 1) /dev/sdc and /dev/sdc used in full with AUFS and 1GB file of COSS in memory 2) /dev/sdc with AUFS and /dev/sdd with COSS (wouldn't be 36 GB too much?) 3) /dev/sdc splitted for COSS and AUFS (if yes, what ratio of splitting), /dev/sdd with AUFS 4) /dev/sdc and /dev/sdd splitted for COSS and AUFS I thought that (1) would be OK. What do you think? What is recommended size of COSS file? -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] IOWAIT and second disk
Matus UHLAR - fantomas wrote: I would use the 4), decrease cache_disk for both disks and create ~1GB file for COSS on both disks. Can you explain why? I am not saying you're wrong, but I want to understand :) -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Stability problems with diskd in Squid 2.6
Anders Nordby wrote: I am using diskd, with 5 cache dirs on each server (each dir containing quite a lot of directories). If I turn off diskd and use plain ufs cache dirs, the instabilitity problem goes away, but then the servers become very slow to respond. :-( Which I/we can not live with. As 2.6 is supposed to be faster than 2.5 because of kqueue, downgrading doesn't seem like a good option either. Funny thing. I had trouble with squid segfaulting in some versions with diskd. Now with 2.5stable13 it works fine. Maybe try this flavour. BTW: does BSD have SMP-aware threads? if yes aufs should perform well (correct me if I am wrong anywhere). -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
[squid-users] CacheMgr: Process Filedescriptor Allocation
Hello! For last two days I had trouble with squid performance. It turns out that some kind of virus was heavily flooding with SYN packets which reduced availible numer of file descriptors and degraded performance I used menu Process Filedescriptor Allocation to finally find that guy. But it would be much easier if there was: a) option to detect IP with most accesses (those SYN do not show in access.log so cachemgr is the only way or using tcpdump) b) at least a way to sort that menu in cachemgr by IP address Maybe you have any other ideas? I know that simples approach would be to limit SYN packets per second but I have no idea of proper limit :) -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] How to customize ERR page with tag for user-agent ?
Alexander Grüner napisał(a): Hello :-) I am running squid-2.5.STABLE9-4.6 on SuSE Linux 9.2 (i586). My squid.conf directs browsers with certain malware user-agents to an error page: ... acl browser-deny browser -i /etc/squid/squid.deny-useragents I can't help you with customisation but I am curious - can we see squid.deny-useragenst contents? :-). Where you get those? -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
[squid-users] Flooding squid
Hello! I have some trouble with new kind of flood targeted at proxy server. One hosts creates thousands of new connections. Is there a way to protect against that at squid level? I would like to avoid doing it with netfilter, because it is hard to guess acceptable limit of connections (browsers tend to open many of them when viewing one page with many pictures, flash, java applets and so on). It is snippet from access.log. 1144674534.008 99296 A.B.C.D TCP_MISS/000 0 GET http://A.B.223.254/ - NONE/- - 1144674534.008 99220 A.B.C.D TCP_MISS/000 0 GET http://A.B.195.224/ - NONE/- - 1144674534.008 99220 A.B.C.D TCP_MISS/000 0 GET http://A.B.209.226/ - NONE/- - 1144674535.013 99967 A.B.C.D TCP_MISS/000 0 GET http://A.B.200.227/ - NONE/- - 1144674535.013 99967 A.B.C.D TCP_MISS/000 0 GET http://A.B.160.207/ - NONE/- - 1144674535.013 99707 A.B.C.D TCP_MISS/000 0 GET http://A.B.235.173/ - NONE/- - 1144674535.022 99967 A.B.C.D TCP_MISS/000 0 GET http://A.B.73.181/ - NONE/- - 1144674535.022 99967 A.B.C.D TCP_MISS/000 0 GET http://A.B.152.222/ - NONE/- - 1144674535.022 99656 A.B.C.D TCP_MISS/000 0 GET http://A.B.92.168/ - NONE/- - 1144674535.022 99397 A.B.C.D TCP_MISS/000 0 GET http://A.B.58.170/ - NONE/- - 1144674535.022 99918 A.B.C.D TCP_MISS/000 0 GET http://A.B.73.28/ - NONE/- - 1144674535.022 99723 A.B.C.D TCP_MISS/000 0 GET http://A.B.120.39/ - NONE/- - 1144674535.022 99731 A.B.C.D TCP_MISS/000 0 GET http://A.B.148.142/ - NONE/- - 1144674535.022 99854 A.B.C.D TCP_MISS/000 0 GET http://A.B.72.163/ - NONE/- - 1144674535.022 99854 A.B.C.D TCP_MISS/000 0 GET http://A.B.225.239/ - NONE/- - 1144674536.005 99964 A.B.C.D TCP_MISS/000 0 GET http://A.B.90.156/ - NONE/- - 1144674536.005 99964 A.B.C.D TCP_MISS/000 0 GET http://A.B.17.97/ - NONE/- - 1144674536.005 99258 A.B.C.D TCP_MISS/000 0 GET http://A.B.91.184/ - NONE/- - 1144674536.005 99938 A.B.C.D TCP_MISS/000 0 GET http://A.B.170.209/ - NONE/- - 1144674536.005 99042 A.B.C.D TCP_MISS/000 0 GET http://A.B.221.13/ - NONE/- - 1144674536.005 99953 A.B.C.D TCP_MISS/000 0 GET http://A.B.147.106/ - NONE/- - 1144674536.005 99898 A.B.C.D TCP_MISS/000 0 GET http://A.B.161.111/ - NONE/- - 1144674536.005 99324 A.B.C.D TCP_MISS/000 0 GET http://A.B.119.65/ - NONE/- - 1144674536.005 99325 A.B.C.D TCP_MISS/000 0 GET http://A.B.115.220/ - NONE/- - 1144674536.005 99387 A.B.C.D TCP_MISS/000 0 GET http://A.B.164.102/ - NONE/- - 1144674536.005 99372 A.B.C.D TCP_MISS/000 0 GET http://A.B.201.135/ - NONE/- - 1144674536.005 99333 A.B.C.D TCP_MISS/000 0 GET http://A.B.52.233/ - NONE/- - 1144674536.005 99362 A.B.C.D TCP_MISS/000 0 GET http://A.B.91.93/ - NONE/- - 1144674536.005 99325 A.B.C.D TCP_MISS/000 0 GET http://A.B.122.22/ - NONE/- - 1144674536.005 99138 A.B.C.D TCP_MISS/000 0 GET http://A.B.10.218/ - NONE/- - -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Flooding squid
Mark Elsen napisał(a): There is a : max_conn parameter, check it out ; in squid.conf.default. Exact name slips me for the moment M. I was thinking about something more sophisticated, because now I have: acl too-many-connections maxconn 500 http_access deny all too-many-connections deny_info ERR_TOO_MANY_CONNECTIONS too-many-connections And it didn't help. Lowering that limit would help, but it will cause problems with people having NAT and few computers behind their router. -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Segmentation fault on x86_64
H napisał(a): I have not read this thread from the beginning but I run since 6 month or so several FreeBSD-amd64 (including X2) servers extremely stable and with real performance advantage in comparism to i386 platforms. It usually happens to me. There are bugs that nobody else notices, because of load, number of clients, or solar radiation. I have no idea why :-). It may also be a small detail, that breaks everything. My coworker noticed that crashes are directly connected with number of viruses hitting proxy. More viruses means much faster crash. -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Segmentation fault on x86_64
Richard Lyons napisał(a): gcc-3.4 is known to generate incorrect code when optimising on x86_64 (http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21804), it may be the cause of the segfault. If removing the optimisations doesn't help you might try gcc-4. Rick. OK it wasn't caused by optimisations. I removed all flags and left only -march=nocona. It seems that thiis is a squid bug, bug on squids bugzilla they ask me to run malloc debuger which I can't do. I am afraid it is good time to go back to 32 bit :\. -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Segmentation fault on x86_64
Michał Margula napisał(a): I am sorry. I don't know why I skipped that. Now I have compiled squid with symbols, enabled core dumps and waiting. Funny thing is that it didn't fail for last 30 hours (usualy it happens much faster!). I hope I am not victim of Murphy Law and it will fail sometime :D. OK. It got sigsegv and I have a ''core'' file. After runing gdb: faramir tmp # gdb /usr/sbin/squid core GNU gdb 6.4 Copyright 2005 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type show copying to see the conditions. There is absolutely no warranty for GDB. Type show warranty for details. This GDB was configured as x86_64-pc-linux-gnu...Using host libthread_db library /lib/tls/libthread_db.so.1. Core was generated by `(squid) -DYC'. Program terminated with signal 11, Segmentation fault. Reading symbols from /lib64/libcrypt.so.1...done. Loaded symbols for /lib/libcrypt.so.1 Reading symbols from /usr/lib64/libssl.so.0.9.7...done. Loaded symbols for /usr/lib/libssl.so.0.9.7 Reading symbols from /usr/lib64/libcrypto.so.0.9.7...done. Loaded symbols for /usr/lib/libcrypto.so.0.9.7 Reading symbols from /lib64/libdl.so.2...done. Loaded symbols for /lib/libdl.so.2 Reading symbols from /lib64/tls/librt.so.1...done. Loaded symbols for /lib/tls/librt.so.1 Reading symbols from /lib64/tls/libpthread.so.0...done. Loaded symbols for /lib/tls/libpthread.so.0 Reading symbols from /lib64/tls/libm.so.6...done. Loaded symbols for /lib/tls/libm.so.6 Reading symbols from /lib64/libresolv.so.2...done. Loaded symbols for /lib/libresolv.so.2 Reading symbols from /lib64/libnsl.so.1...done. Loaded symbols for /lib/libnsl.so.1 Reading symbols from /lib64/tls/libc.so.6...done. Loaded symbols for /lib/tls/libc.so.6 Reading symbols from /lib64/ld-linux-x86-64.so.2...done. Loaded symbols for /lib64/ld-linux-x86-64.so.2 Reading symbols from /lib64/libnss_compat.so.2...done. Loaded symbols for /lib/libnss_compat.so.2 Reading symbols from /lib64/libnss_nis.so.2...done. Loaded symbols for /lib/libnss_nis.so.2 Reading symbols from /lib64/libnss_files.so.2...done. Loaded symbols for /lib/libnss_files.so.2 #0 0x2b798b16 in malloc_usable_size () from /lib/tls/libc.so.6 (gdb) where #0 0x2b798b16 in malloc_usable_size () from /lib/tls/libc.so.6 #1 0x2b799848 in free () from /lib/tls/libc.so.6 #2 0x2b79add0 in calloc () from /lib/tls/libc.so.6 #3 0x004a1c3c in xcalloc () #4 0x004594b6 in memBufGrow () #5 0x00446d45 in httpStart () #6 0x004369a0 in fwdConnectDone () #7 0x004371a1 in fwdConnectStart () #8 0x00461c3a in peerSelectFoo () #9 0x00436587 in fwdStart () #10 0x00429ea6 in clientCacheHit () #11 0x0047561d in storeClientReadHeader () #12 0x00485aa7 in storeDiskdDirCallback () #13 0x00477d6d in storeDirCallback () #14 0x0048a06b in storeDiskdSend () #15 0x0048a358 in storeDiskdWrite () #16 0x00479d37 in storeSwapOut () #17 0x00445c2f in httpReadReply () #18 0x0042e524 in comm_poll () #19 0x00456fa5 in main () Is it helpful in any way? -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Segmentation fault on x86_64
Serassio Guido napisał(a): Hi, At 01.36 10/12/2005, =?ISO-8859-2?Q?Micha=B3_Margula?= wrote: Hello! I was trying to find solution for my problem for a long long time, but I had no luck. Everything started when I installed Gentoo with Squid on 64 bit architecture. It is dual Xeon 3.4 with 2 GB of RAM. My CFLAGS are -O3 -march=nocona -pipe -fomit-frame-pointer. It didn't happen on 32 bit architecture. What gcc version ? A first try could be don't use any code optimization, many times they generate more problems then benefits. gcc version 3.4.4 (Gentoo 3.4.4-r1, ssp-3.4.4-1.0, pie-8.7.8). Those optimizations didn't cause problems earlier, I will try remove them if I can't find anything in core dump. What happens in cache.log file ? 2005/12/10 09:03:33| Request header is too large (21900 bytes) 2005/12/10 09:03:33| Config 'request_header_max_size'= 20480 bytes. 2005/12/10 09:03:33| Request header is too large (21900 bytes) 2005/12/10 09:03:33| Config 'request_header_max_size'= 20480 bytes. 2005/12/10 09:03:34| Request header is too large (23360 bytes) 2005/12/10 09:03:34| Config 'request_header_max_size'= 20480 bytes. 2005/12/10 09:03:46| Starting Squid Cache version 2.5.STABLE12 for x86_64-pc-linux-gnu... So nothing usual. I get plenty of WARNING: failed to unpack meta data, but those are cause by segfaults of squid. I would appreciate any help, maybe somebody had same issue, or some kind of Squid-debugging-HOWTO. I can try that also. What you are asking about was always in Squid's FAQ: http://www.squid-cache.org/Doc/FAQ/FAQ-11.html#ss11.19 http://www.squid-cache.org/Doc/FAQ/FAQ-11.html#ss11.20 I am sorry. I don't know why I skipped that. Now I have compiled squid with symbols, enabled core dumps and waiting. Funny thing is that it didn't fail for last 30 hours (usualy it happens much faster!). I hope I am not victim of Murphy Law and it will fail sometime :D. -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
Re: [squid-users] Segmentation fault on x86_64
Richard Lyons napisał(a): gcc-3.4 is known to generate incorrect code when optimising on x86_64 (http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21804), it may be the cause of the segfault. If removing the optimisations doesn't help you might try gcc-4. Rick. Funny thing is that now squid works fine (I recompiled it with debugging symbols). If it is fine, I will try compilation without omptimising and post about results. Thanks -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]
[squid-users] Segmentation fault on x86_64
Hello! I was trying to find solution for my problem for a long long time, but I had no luck. Everything started when I installed Gentoo with Squid on 64 bit architecture. It is dual Xeon 3.4 with 2 GB of RAM. My CFLAGS are -O3 -march=nocona -pipe -fomit-frame-pointer. It didn't happen on 32 bit architecture. Problem is that few hours after starting squid it starts to issue segmentation faults. It is snippet of my logs: Dec 9 13:00:58 faramir squid[18109]: segfault at 0001015cc190 rip 2b798b16 rsp 7fc3c4f0 error 4 Dec 9 13:25:46 faramir squid[18315]: segfault at 000100d161f0 rip 2b798b16 rsp 7fd8f2f0 error 4 Dec 9 13:47:45 faramir squid[18601]: segfault at 000101ad26b8 rip 2b7995a1 rsp 7ff5bab0 error 4 Dec 9 14:11:10 faramir squid[18856]: segfault at 000100bd8000 rip 2b798b16 rsp 7fd3f370 error 4 Dec 9 14:34:50 faramir squid[19151]: segfault at 000101701d80 rip 2b798b16 rsp 7f962460 error 4 Dec 9 14:55:35 faramir squid[19417]: segfault at 0001019c5930 rip 2b798b16 rsp 7f80a8e0 error 4 Dec 9 15:22:28 faramir squid[19682]: segfault at 00010010 rip 2b798b16 rsp 7f8a60b0 error 4 Dec 9 15:43:24 faramir squid[20001]: segfault at 000101ade030 rip 2b798b16 rsp 7fde07e0 error 4 Dec 9 16:07:08 faramir squid[20252]: segfault at 0001018a1ec0 rip 2b798b16 rsp 7f80b170 error 4 Dec 9 16:37:23 faramir squid[20536]: segfault at 00010010 rip 2b798b16 rsp 7fea0800 error 4 Dec 9 17:03:36 faramir squid[20878]: segfault at 00010aacfd10 rip 2b798b16 rsp 7f83a370 error 4 Dec 9 17:30:09 faramir squid[21189]: segfault at 0001013c3590 rip 2b798b16 rsp 7fa677f0 error 4 Dec 9 17:57:08 faramir squid[21516]: segfault at 000100d524c0 rip 2b798b16 rsp 7fb5d420 error 4 Dec 9 18:24:16 faramir squid[21816]: segfault at 0001014d0bf0 rip 2b798b16 rsp 7fdf3b60 error 4 Dec 9 18:49:24 faramir squid[22131]: segfault at 000101390710 rip 2b798b16 rsp 7ff6f960 error 4 Dec 9 19:16:18 faramir squid[22420]: segfault at 00010010 rip 2b798b16 rsp 7f8ccdf0 error 4 Dec 9 19:45:15 faramir squid[22791]: segfault at 000101357630 rip 2b798b16 rsp 7f9286a0 error 4 Dec 9 20:11:12 faramir squid[23136]: segfault at 0001014d8610 rip 2b798b16 rsp 7ff0ad60 error 4 Dec 9 20:22:43 faramir squid[23494]: segfault at 00010123d540 rip 2b798b16 rsp 7fa35700 error 4 Dec 9 20:48:16 faramir squid[23648]: segfault at 000101dea870 rip 2b798b16 rsp 7feb17a0 error 4 Dec 9 21:07:01 faramir squid[23939]: segfault at 0001016aef40 rip 2b798b16 rsp 7fd6aab0 error 4 Dec 9 21:41:41 faramir squid[24184]: segfault at 0001015a8570 rip 2b798b16 rsp 7fbd03a0 error 4 Dec 9 22:14:04 faramir squid[24580]: segfault at 00010f49d1e0 rip 2b798b16 rsp 7fb96fc0 error 4 Dec 9 22:29:50 faramir squid[24943]: segfault at 00010157a410 rip 2b798b16 rsp 7fd6e990 error 4 Dec 9 22:43:39 faramir squid[25135]: segfault at 000100f01870 rip 2b798b16 rsp 7fb0d630 error 4 Dec 9 23:11:11 faramir squid[25330]: segfault at 0001020292d0 rip 2b798b16 rsp 7fedadc0 error 4 Dec 9 23:23:45 faramir squid[25663]: segfault at 000101323b80 rip 2b798b16 rsp 7fb9c0c0 error 4 Dec 9 23:42:52 faramir squid[25821]: segfault at 0001017e2c90 rip 2b798b16 rsp 7fe44c00 error 4 And so on. Always it is error 4, at address changes and rip is almost same - 2b798b16 or 2b7995a1 in last week. I went trough kernel sources and found that error 4 means that segfault was caused by user space app when trying to read page that wasn't found. I am using squid-2.5.STABLE12. It happens also with STABLE10 and STABLE11. Funny thing is that squid is still runing after segfault. When it occurs I receive Connection refused from squid daemon, and then it works fine. But those errors are annoying my customers. At http://alchemyx.uznam.net.pl/squid/ you can find my mrtg graphs from data collected via SNMP. They are really strange. I would appreciate any help, maybe somebody had same issue, or some kind of Squid-debugging-HOWTO. I can try that also. Thank you in advance and Merry Christmas (soon!) -- Michał Margula, [EMAIL PROTECTED], http://alchemyx.uznam.net.pl/ W życiu piękne są tylko chwile [Ryszard Riedel]