[squid-users] squid with 16384 descriptors still running out.
Hi, I haven't posted in a while but I am seeing an old problem reoccur. I am getting the "Your Cache is Running out of FileDescriptors" error again. My platform is RHEL, Stock RHEL kernel, 3.1Gz P4, 1 gig of RAM. Squid stable9. Any danger in doubling the FD again? to 32768 ?? Thanks- __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
[squid-users] Testing cacheability of objects
Hi, I've been finding some docs online that suggest that URLS that contain ? are not cacheable. I think that this was once, perhaps long ago, the default behavior but is it still the case?? If so what does one do to get Squid to chache an object where the url contains a ? Anyone know if there's a good test for cacheability of objects?? I found this one.. http://www.ircache.net/cgi-bin/cacheability.py But it may be too old to be considered usefull. Thanks- __ Celebrate Yahoo!'s 10th Birthday! Yahoo! Netrospective: 100 Moments of the Web http://birthday.yahoo.com/netrospective/
[squid-users] ideas for getting this to work with squid
Hi, My setup is RHEL-3.1 and Squid-2.5-STABLE9+collapsed_forwarding turned on. I am trying to get squid to work in accelerator mode to provide relief to a backend server. The clients are set up so they make requests to the squid box but if squid doesn't have the object in cache, clients go to my origin server,which is running tomcat. This can mean several hundred clients going to the origin server at the same time for the same object. A sad situation for the origin server. I would expect that once the object is received from the origin server it gets cached, but I am seeing that its not being cached. I've verified that the object is cachable. When several requests arrive at the same time I can live with one or two requests going to the origin server, but I ve got to stop each and every one of the requests from continuing to pour into the origin server. How do i get squid to do that?? -any help / pointers / directions to a cliff or bridge appreciated. -RAy __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
[squid-users] help with accelerator mode
Hi, I have some lingering questions from last week that are, well, still with me. So I am hoping to hear from anyone with experience running squid in accelerator mode, having the collapsed_forwarding would be great too. As in my previous posts my setup is RHEL-3.1 and Squid-2.5-STABLE9+collapsed_forwarding turned on. I am trying to get squid to work in accelerator mode to provide relief to a backend server. The clients are set up to point to the squid box but when the cache doesn't have the object in cache, clients go to my origin server. This can mean several hundred clients going to the origin server at the same time for the same object. A sad situation for the origin server. Once the object is received by the origin server its not being cached. I've verified that the object is cachable. When several requests arrive at the same time I can live with one or two going to the origin server but I ve got to stop each and every one of the requests to continue poring into the origin server. How do i get squid to do that?? Thanks for any help, -RAy __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
[squid-users] Acclerator mode
Hi all, Using RHEL-3.1 x86, Squid-STABLE9 + Collapsed_Forward patch. I am try to set up squid as a surrogate to my back end server which has an http service running on port 44003. What I' am seeing when multiple requests for the same URI comes in that rather than allow just one or two of the requests to go to the backend. So, using a separate box, I setup multiple wgets (say 20) to request the same URI ... each request initiates a session to the back end. :( Any thoughts about what is wrong with my test or the config or hints about what folks usually goof up when deploying squid as in Accelerator mode. Thanks in advance for any help/suggestions, Ray Here is my squid.conf file ### ### ### http_port 32011 http_port 21100 http_port 44003 httpd_accel_host mybackendexampleserver.com httpd_accel_port 0 httpd_accel_uses_host_header on httpd_accel_with_proxy on # # icp_port 0 cache_mem 128 MB cache_swap_low 90 cache_swap_high 95 maximum_object_size 1024 MB #maximum_object_size_in_memory 100 MB maximum_object_size_in_memory 8 MB ipcache_size 1024 ipcache_low 90 ipcache_high 95 cache_replacement_policy lru memory_replacement_policy lru # # cache_dir ufs /opt/squid/cache 16384 16 256 cache_access_log /opt/log/squid/access.log cache_log /opt/log/squid/cache.log mime_table /opt/squid/etc/mime.conf pid_filename /var/run/squid.pid ##refresh_pattern . 0 20% 10080 refresh_pattern . 10 20% 10080 # # #redirect_rewrites_host_header off #Tried but no differnce # collapsed_forwarding on refresh_stale_hit 0 seconds acl ALL src 0/0 #acl ProxyUsers src 192.168.10.0/24 #acl ProxyUsers src 192.168.0.0/24 acl ProxyUsers src 192.168.0.0/17 acl TheOriginServer dstdomain mybackendexampleserver.com http_access allow ProxyUsers http_access allow TheOriginServer http_access deny ALL acl manager proto cache_object acl localhost src 127.0.0.1/255.255.255.255 acl SSL_ports port 443 563 acl Safe_ports port 80 acl Safe_ports port 443 563 acl Safe_ports port 1025-65535 acl CONNECT method CONNECT acl snmp snmp_community COMTYPASS acl dstbackenddom dstdomain .backenddom.com acl backend-svc dstdomain .backend-svc.com acl dstpriv dstdomain .backendsvc-priv.com http_access allow localhost http_access allow manager localhost http_access allow dstbackenddom http_access allow dstbackend-svc http_access allow dstpriv http_access allow snmp !localhost http_access deny !Safe_ports http_access deny CONNECT !SSL_ports http_access deny all cache_mgr [EMAIL PROTECTED] cache_effective_user squid cache_effective_group squid logfile_rotate 10 cachemgr_passwd PASSWD shutdown cachemgr_passwd PASSWD info stats/objects cachemgr_passwd disable all icon_directory /opt/squid/share/icons error_directory /opt/squid/share/errors/English snmp_port 3401 snmp_access allow snmp snmp_access deny all uri_whitespace strip strip_query_terms off cache_store_log /opt/log/squid/storage.log no_cache deny CONNECT emulate_httpd_log on __ Do you Yahoo!? Take Yahoo! Mail with you! Get it on your mobile phone. http://mobile.yahoo.com/maildemo
[squid-users] Understanding the stats.
Hi, Looking at some stats from Cache Manager. I originally thought there were just hits or misses, is a Near Hit still a miss? I am hoping for a deeper explaination of Near Hits. Also, can you guys tell me, but doesn't this look like a poorly performing squid server?? It's for approx 30Hours. Connection information for squid: Number of clients accessing cache: 5384 Number of HTTP requests received: 7124911 Number of ICP messages received:0 Number of ICP messages sent:0 Number of queued ICP replies: 0 Request failure ratio: 0.00 Average HTTP requests per minute since start: 3641.7 Average ICP messages per minute since start: 0.0 Select loop called: 26095903 times, 4.498 ms avg Cache information for squid: Request Hit Ratios: 5min: 99.1%, 60min: 98.5% Byte Hit Ratios:5min: 99.0%, 60min: 98.8% Request Memory Hit Ratios: 5min: 78.4%, 60min: 91.2% Request Disk Hit Ratios:5min: 4.5%, 60min: 4.1% Storage Swap size: 802820 KB Storage Mem size: 90464 KB Mean Object Size: 609.12 KB Requests given to unlinkd: 0 Median Service Times (seconds) 5 min60 min: HTTP Requests (All): 0.00678 0.00379 Cache Misses: 0.03241 0.04776 Cache Hits:0.00463 0.00379 Near Hits:169.11253 169.11253 Not-Modified Replies: 0.0 0.0 DNS Lookups: 0.0 0.05078 ICP Queries: 0.0 0.0 Resource usage for squid: UP Time:117388.930 seconds CPU Time: 54245.630 seconds CPU Usage: 46.21% CPU Usage, 5 minute avg:42.93% CPU Usage, 60 minute avg: 38.88% Process Data Segment Size via sbrk(): 179290 KB Maximum Resident Size: 0 KB Page faults with physical i/o: 52592 __ Do you Yahoo!? Yahoo! Mail - 250MB free storage. Do more. Manage less. http://info.mail.yahoo.com/mail_250
RE: [squid-users] What did I do to make squid behace this way??
> Does it work when the collapse forwarding patch is >not used ? > >M The behavior occurs even in a plan vanilla install of squid. So regardless of the patch. Thanks- __ Do you Yahoo!? The all-new My Yahoo! - What will yours do? http://my.yahoo.com
[squid-users] What did I do to make squid behace this way??
Hi, My setup- squid-stable8-20050218 w/collapsed_forwarding patch. RHEL-3.1 on 3.6Ghz w/1Gig of RAM. SW Raid. I've been trying to understand how squid works for the past week or so and I'am looking for what causes the following behavior when squid receives a request to fetch a URI not associated with my network: Squid stops w/SEGV message 2005/02/18 20:54:34| storeLateRelease: released 0 objects FATAL: Received Segment Violation...dying. 2005/02/18 21:10:56| storeDirWriteCleanLogs: Starting... 2005/02/18 21:10:56| WARNING: Closing open FD 12 Any insights or pointers much appreciatied, Ray __ Do you Yahoo!? Yahoo! Mail - Find what you need with new enhanced search. http://info.mail.yahoo.com/mail_250
[squid-users] build no longer fails after system upgrade
Hi, My original post on this matter lacked critical infosorry bout that. I've been able to find where my mistake was, it was not a problem with the patch/or my update. Thanks everyone ! __ Do you Yahoo!? Yahoo! Mail - 250MB free storage. Do more. Manage less. http://info.mail.yahoo.com/mail_250
[squid-users] build failing after sys upgrade
Hi, I am sure that my problem is a direct result of a recent system update that ran yesterday. I kind of thought my kerberos needed updating but doing so didn't make a difference. A vinilla squid build works but when I apply the patch for collapsed_forwarding I get the following errors: gcc -DHAVE_CONFIG_H -DDEFAULT_CONFIG_FILE=\"/opt/squid/etc/squid.conf\" -I. -I. -I../include -I. -I. -I../include -I../include -I/usr/kerberos/include -g -O2 -Wall -D_REENTRANT -c `test -f client_side.c || echo './'`client_side.c client_side.c: In function `clientProcessExpired': client_side.c:437: `squid' undeclared (first use in this function) client_side.c:437: (Each undeclared identifier is reported only once client_side.c:437: for each function it appears in.) client_side.c:437: too many decimal points in floating constant client_side.c:455: too many decimal points in floating constant client_side.c: In function `clientCacheHit': client_side.c:1637: `squid' undeclared (first use in this function) client_side.c:1637: too many decimal points in floating constant client_side.c: In function `clientProcessMiss': client_side.c:2548: `squid' undeclared (first use in this function) client_side.c:2548: too many decimal points in floating constant make[3]: *** [client_side.o] Error 1 Looking for suggestions.. __ Do you Yahoo!? Take Yahoo! Mail with you! Get it on your mobile phone. http://mobile.yahoo.com/maildemo
[squid-users] Increasing the File Descriptors
I am getting errors about running out of FD and I checked d the FAQ and found a couple pointers for doing a ulimit before running ./config and editing the /usr/include/bit/types.h file. I just wanted to ask if I should also increase the FD_SETSIZE ?? Probably not, but here is that part of ./config's output. checking Default FD_SETSIZE value... 1024 checking Maximum number of filedescriptors we can open... 16384 Sorry if this seems like an obvious question but What caching situation calls for a bigger FD_SIZE ?? Thanks- Ray __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
[squid-users] Is this the correct patch to use?
Hi, I am running squid-2.5.stable-7 on RHEL-3. I am seeing multiple requests for the same web object and its creating a bottle-neck. The original request appears to never be cachable. After some reading and googling I'am thinking that the collapse_forwarding patch will help me. What else should I check?? -- Thanks in advance __ Do you Yahoo!? The all-new My Yahoo! - What will yours do? http://my.yahoo.com