Hi Mr. Jeffreis, 2011/1/7 Amos Jeffries <squid3@xxxxxxxxxxxxx>: > On 08/01/11 00:44, Robert Pipca wrote: >> >> Hi Mr. Jeffreis, >> >> 2011/1/7 Robert Pipca<robertpipca@xxxxxxxxx>: >>> >>> 2011/1/7 Amos Jeffries<squid3@xxxxxxxxxxxxx>: >>>> >>>> Your config shows ~69 GB of small files. Each cache_dir has a maximum >>>> count >>>> of 2^31 files. It looks like that file count is being exceeded and the >>>> overflow handling is broken. >> >> I looked at one server showing this bug, and mgr:info gives me this: >> >> Internal Data Structures: >> 8152065 StoreEntries >> 271 StoreEntries with MemObjects >> 27 Hot Object Cache Items >> 8151819 on-disk objects >> >> Since 2^31 is around 2+ billion files, it doesn't seem that maximum >> count of files is being reached on all cache_dirs... > > Thats a good sign in a way. The fix looks like it will be a patch, just a > matter of finding it. That's good :) We really need to use cache for bandwidth saving, and besides this, squid is working like a charm. >> >> Do you have any other ideas as to why this bug keeps happening? >> >> I don't know how to reproduce it, but I have around 10 squid machines >> on different operations of our ISP showing this issue, all with the >> same coss setup. >> >> Cheers, >> >> - Robert. > > You will have to contact Adrian about the Lusca side of things. That is his > private commercial fork. > > Henrik may be able to help with the squid-2 bits. I have 10 different operations (cities) of the infra-structure. I put Lusca in 5 of them, and squid-stable9 on the other 5. This assertion failed occurs on all of them. So, Mr. Henrik, can you help hunt down this bug on squid like mr. Jeffreis suggested? Thanks. Cheers, - Robert