Re: [squid-users] worker per cache_dir
On 7/06/2015 1:40 a.m., Marcel Fossua wrote: ls -la /dev/shm/ total 0 drwxrwxrwt 2 proxy proxy 40 Jun 6 14:56 . drwxr-xr-x 14 root root 3920 Jun 6 02:07 .. That should be root:root I believe. Its an operating system component. Though I'm not quite as certain that it will work any different once you fix that. Maybe, maybe not. Amos ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
ls -la /dev/shm/ total 0 drwxrwxrwt 2 proxy proxy 40 Jun 6 14:56 . drwxr-xr-x 14 root root 3920 Jun 6 02:07 .. -- View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/worker-per-cache-dir-tp4671510p4671576.html Sent from the Squid - Users mailing list archive at Nabble.com. ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
Hi Amos any hints to help me with this issue I not able to set more than on cache_dir per worker because it's alway end up with errors (but no errors appears if only one cache_dir is set per worker) 2015/06/06 01:36:09 kid7| Starting eCAP service: ecap://www.vigos.com/ecap_gzip 2015/06/06 01:36:09 kid7| commBind: Cannot bind socket FD 9 to [::]: (13) Permission denied 2015/06/06 01:36:09 kid6| commBind: Cannot bind socket FD 21 to [::]: (13) Permission denied 2015/06/06 01:36:09 kid6| Store rebuilding is 0.00% complete FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. CPU Usage: 0.016 seconds = 0.016 user + 0.000 sys Maximum Resident Size: 31344 KB Page faults with physical i/o: 0 2015/06/06 01:36:14 kid6| Store rebuilding is 43.40% complete FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. CPU Usage: 0.016 seconds = 0.016 user + 0.000 sys Maximum Resident Size: 31328 KB Page faults with physical i/o: 0 FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory one cache_dir/worker ** cache_dir rock /var/spool/squid3/cache10 46 min-size=1 max-size=31000 max-swap-rate=200 swap-timeout=300 # a 200GB x 8 caches of large ( over 32KB) objects per-worker if ${process_number} = 1 cache_dir aufs /cache0 350 32 256 min-size=31001 max-size=104857600 endif if ${process_number} = 2 cache_dir aufs /cache1 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 3 cache_dir aufs /cache2 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 4 cache_dir aufs /cache3 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 5 cache_dir aufs /cache4 350 32 256 min-size=31001 max-size=1048576000 endif 2 cache_dir/worker * #if ${process_number} = 4 #cache_dir aufs /cache6/${process_number} 350 32 256 min-size=31001 max-size=1048576000 #cache_dir aufs /cache7/${process_number} 350 32 256 min-size=31001 max-size=1048576000 #endif #if ${process_number} = 5 #cache_dir aufs /cache8/${process_number} 350 32 256 min-size=31001 max-size=1048576000 #cache_dir aufs /cache9/${process_number} 350 32 256 min-size=31001 max-size=1048576000 #endif Thanks -- View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/worker-per-cache-dir-tp4671510p4671571.html Sent from the Squid - Users mailing list archive at Nabble.com. ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
On 6/06/2015 12:26 p.m., Marcel Fossua wrote: Hi Amos any hints to help me with this issue I not able to set more than on cache_dir per worker because it's alway end up with errors (but no errors appears if only one cache_dir is set per worker) 2015/06/06 01:36:09 kid7| Starting eCAP service: ecap://www.vigos.com/ecap_gzip 2015/06/06 01:36:09 kid7| commBind: Cannot bind socket FD 9 to [::]: (13) Permission denied 2015/06/06 01:36:09 kid6| commBind: Cannot bind socket FD 21 to [::]: (13) Permission denied These are a Squid process not being able to open /dev/shm UDS sockets. 2015/06/06 01:36:09 kid6| Store rebuilding is 0.00% complete FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. CPU Usage: 0.016 seconds = 0.016 user + 0.000 sys Maximum Resident Size: 31344 KB Page faults with physical i/o: 0 2015/06/06 01:36:14 kid6| Store rebuilding is 43.40% complete FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. CPU Usage: 0.016 seconds = 0.016 user + 0.000 sys Maximum Resident Size: 31328 KB Page faults with physical i/o: 0 FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory Squid Cache (Version 3.5.5): Terminated abnormally. FATAL: Ipc::Mem::Segment::open failed to shm_open(/squid-cf__metadata.shm): (2) No such file or directory These ones are the result of /dev/shm UDS sockets not existing. Check the /dev/shm has write permission for everyone. And that Squid is fully shutdown and restarted (not just reconfigured). And that no other Squid process is already running - that includes a previous squid -z. Amos ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
On 5/06/2015 1:27 a.m., Marcel Fossua wrote: Cool Thanks but I have an error while doing that maybe it could be the HDD size By the way Amos what could you suggest me to handle disks I have a jbod with 15 disks (4TB) each I read on of your comment stipulating to set a cache_dir per drive (o I'm totaly wrong) That is correct, only place one *AUFS/UFS/diskd cache_dir* per physical drive disk. But that is because Squid will overload the HDD controller with seek and I/O collisions when 2 or more are stored together. with this worker/disk distribution I end up with error Aha, which you have forgotten to copy so we can see exactly what you are going on about. If I assume, I would guess it is more likely a 32-bit rounding error calculating how big in Bytes the cache is or the max-size=N object parameter. Though the total size should be in 64-bit math by now. Amos ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
Hi Amos this is the output error *2015/06/04 16:18:22 kid1| Logfile: opening log stdio:/var/log/squid/error.log FATAL: xcalloc: Unable to allocate 18446744073689603781 blocks of 1 bytes! 2015/06/04 16:18:37 kid4| Logfile: opening log stdio:/var/log/squid/error.log FATAL: xcalloc: Unable to allocate 18446744073689603781 blocks of 1 bytes! .. * El 4/6/15 a las 15:11, Amos Jeffries escribió: On 5/06/2015 1:27 a.m., Marcel Fossua wrote: Cool Thanks but I have an error while doing that maybe it could be the HDD size By the way Amos what could you suggest me to handle disks I have a jbod with 15 disks (4TB) each I read on of your comment stipulating to set a cache_dir per drive (o I'm totaly wrong) That is correct, only place one *AUFS/UFS/diskd cache_dir* per physical drive disk. But that is because Squid will overload the HDD controller with seek and I/O collisions when 2 or more are stored together. with this worker/disk distribution I end up with error Aha, which you have forgotten to copy so we can see exactly what you are going on about. If I assume, I would guess it is more likely a 32-bit rounding error calculating how big in Bytes the cache is or the max-size=N object parameter. Though the total size should be in 64-bit math by now. Amos ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users -- Fossua-vcard Marcel Fossua Unix/Linux Network Administrator Tel: 0240 99448 www.guineanet.net http://www.guineanet.net/ www.familyfossua.com http://www.familyfossua.com ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
Cool Thanks but I have an error while doing that maybe it could be the HDD size By the way Amos what could you suggest me to handle disks I have a jbod with 15 disks (4TB) each I read on of your comment stipulating to set a cache_dir per drive (o I'm totaly wrong) with this worker/disk distribution I end up with error cache_dir rock /cache16 50 min-size=1 max-size=31000 max-swap-rate=200 swap-timeout=300 if ${process_number} = 1 cache_dir aufs /cache0 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache1 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 2 cache_dir aufs /cache2 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache3 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 3 cache_dir aufs /cache4 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache5 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 4 cache_dir aufs /cache6 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache7 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 5 cache_dir aufs /cache8 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache9 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 6 cache_dir aufs /cache10 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache11 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 7 cache_dir aufs /cache12 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache13 350 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 8 cache_dir aufs /cache14 350 32 256 min-size=31001 max-size=1048576000 cache_dir aufs /cache15 350 32 256 min-size=31001 max-size=1048576000 endif -- View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/worker-per-cache-dir-tp4671510p4671541.html Sent from the Squid - Users mailing list archive at Nabble.com. ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
Re: [squid-users] worker per cache_dir
On 4/06/2015 5:18 a.m., Marcel Fossua wrote: Hi All hope someone can give me a way to accomplish what I have in mind I have 3.5.5 running with 9 workers active (the best numbers a get without errors) so I just set 1 worker per disk as schema below but I have a jbod with lot of disk I would like to add on squid.conf obviously the idea of 1 worker /disk is not longer a good deal then what is the best config to make let say worker 1 deal with several cache (3 or 4 for the occurence) so I could set 3 o 4 cache_dir per worker. Thanks Sure, inside the worker if...endif block list the other cache_dir you want it to deal with. Each worker can have up to 63. So long as the AUFS cache_dir are only used by one worker it will work fine. It is the other way around which is broken - multiple worker using one AUFS cache_dir. Amos ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users
[squid-users] worker per cache_dir
Hi All hope someone can give me a way to accomplish what I have in mind I have 3.5.5 running with 9 workers active (the best numbers a get without errors) so I just set 1 worker per disk as schema below but I have a jbod with lot of disk I would like to add on squid.conf obviously the idea of 1 worker /disk is not longer a good deal then what is the best config to make let say worker 1 deal with several cache (3 or 4 for the occurence) so I could set 3 o 4 cache_dir per worker. Thanks /cache_dir rock /cache1 46 min-size=1 max-size=31000 max-swap-rate=200 swap-timeout=300 # 200GB x 8 caches of large ( over 32KB) objects per-worker if ${process_number} = 1 cache_dir aufs /cache2 275000 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 2 cache_dir aufs /cache3 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 3 cache_dir aufs /cache4 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 4 cache_dir aufs /cache5 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 5 cache_dir aufs /cache6 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 6 cache_dir aufs /cache7 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 7 cache_dir aufs /cache8 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 8 cache_dir aufs /cache9 19 32 256 min-size=31001 max-size=1048576000 endif if ${process_number} = 9 cache_dir aufs /cache10 19 32 256 min-size=31001 max-size=1048576000 endif/ -- View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/worker-per-cache-dir-tp4671510.html Sent from the Squid - Users mailing list archive at Nabble.com. ___ squid-users mailing list squid-users@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-users