[squid-users] can't use diskd (on FreeBSD)

Hello!

I'm trying to see if using the diskd will help me, but squid dies
promptly after serving a few requests with:

        assertion failed: store_io_diskd.c:492: "++send_errors < 100"

I'm seeing this on two machines, which are identical in everything
except that on one there are 4 independented cache_dir -- on for each
drive, while on the other all 4 partitions are striped into one.

The ufs method works fine on both (it is, evidently, better to use the
separate cache_dirs, though), but the diskd method fails on both :(

More from the squid-cache.log (date removed for neatness):

        [...]
        08:56:44| file_open: error opening file /squid/netdb_state: (2) No such file or directory
        [...]
        08:57:06| storeClientCopy: C0F372EA88D5A490A6FC3DB26A1E4EFA, seen 49152, want 49152, size 4096, cb 0x8060514, cbdata 0x8e93400
        08:57:06| cbdataLock: 0x8e91340
        08:57:06| storeClientCopy2: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| storeClientCopy3: Copying from memory
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| cbdataLock: 0x8e93400
        08:57:06| cbdataUnlock: 0x8e91340
        08:57:06| cbdataUnlock: 0x8e93400
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| storeClientCopy: C0F372EA88D5A490A6FC3DB26A1E4EFA, seen 53248, want 53248, size 4096, cb 0x8060514, cbdata 0x8e93400
        08:57:06| cbdataLock: 0x8e91340
        08:57:06| storeClientCopy2: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| storeClientCopy3: Copying from memory
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| cbdataLock: 0x8e93400
        08:57:06| cbdataUnlock: 0x8e91340
        08:57:06| cbdataUnlock: 0x8e93400
        08:57:06| commSetTimeout: FD 26 timeout 900
        08:57:06| InvokeHandlers: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| InvokeHandlers: checking client #0
        08:57:06| storeSwapMetaBuild: http://pie.privatelabs.com:8000/IPC/;URL:http://images.privatelabs.com/Images/BSD_Chick.jpg;cache:;monochrome:;rotate:12
        08:57:06| cbdataAdd: 0x8e8afc0
        08:57:06| storeCreate: Selected dir '0' for obj size '-1'
        08:57:06| storeDiskdCreate: fileno 00000062
        08:57:06| cbdataAdd: 0x8ecad40
        08:57:06| cbdataLock: 0x8e8afc0
        08:57:06| cbdataLock: 0x8ecad40
        08:57:06| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
        08:57:06| cbdataUnlock: 0x8ecad40
        08:57:06| storeDiskdSend OPEN: (35) Resource temporarily unavailable
        08:57:06| cbdataUnlock: 0x8e8afc0
        08:57:06| cbdataFree: 0x8ecad40
        08:57:06| cbdataReallyFree: Freeing 0x8ecad40
        08:57:06| cbdataFree: 0x8e8afc0
        08:57:06| cbdataReallyFree: Freeing 0x8e8afc0
        08:57:06| httpPconnTransferDone: FD 26
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| storeClientCopy: C0F372EA88D5A490A6FC3DB26A1E4EFA, seen 57344, want 57344, size 4096, cb 0x8060514, cbdata 0x8e93400
        08:57:06| cbdataLock: 0x8e91340
        08:57:06| storeClientCopy2: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| storeClientCopy3: Copying from memory
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| cbdataLock: 0x8e93400
        08:57:06| cbdataUnlock: 0x8e91340
        08:57:06| cbdataUnlock: 0x8e93400
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| storeClientCopy: C0F372EA88D5A490A6FC3DB26A1E4EFA, seen 61440, want 61440, size 4096, cb 0x8060514, cbdata 0x8e93400
        08:57:06| cbdataLock: 0x8e91340
        08:57:06| storeClientCopy2: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| storeClientCopy3: Copying from memory
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| cbdataLock: 0x8e93400
        08:57:06| cbdataUnlock: 0x8e91340
        08:57:06| cbdataUnlock: 0x8e93400
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| storeClientCopy: C0F372EA88D5A490A6FC3DB26A1E4EFA, seen 65536, want 65536, size 4096, cb 0x8060514, cbdata 0x8e93400
        08:57:06| cbdataLock: 0x8e91340
        08:57:06| storeClientCopy2: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| storeClientCopy3: Copying from memory
        08:57:06| cbdataValid: 0x8e93400
        08:57:06| cbdataLock: 0x8e93400
        08:57:06| cbdataUnlock: 0x8e91340
        08:57:06| cbdataUnlock: 0x8e93400
        08:57:06| commSetTimeout: FD 26 timeout 900
        08:57:06| InvokeHandlers: C0F372EA88D5A490A6FC3DB26A1E4EFA
        08:57:06| InvokeHandlers: checking client #0
        08:57:06| storeSwapMetaBuild: http://pie.privatelabs.com:8000/IPC/;URL:http://images.privatelabs.com/Images/BSD_Chick.jpg;cache:;monochrome:;rotate:12
        08:57:06| cbdataAdd: 0x8f53030
        08:57:06| storeCreate: Selected dir '0' for obj size '-1'
        08:57:06| storeDiskdCreate: fileno 00000063
        08:57:06| cbdataAdd: 0x8ecad40
        08:57:06| cbdataLock: 0x8f53030
        08:57:06| cbdataLock: 0x8ecad40
        08:57:06| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
        08:57:06| cbdataUnlock: 0x8ecad40
        08:57:06| assertion failed: store_io_diskd.c:492: "++send_errors < 100"
        [boom -- this is where squid aborts on signal 6]

Any ideas? I'm using the squid-2.4.STABLE1 with the following patches:

        squid-2.4.stable1-wrong_sign_on_timestamp_check.patch
        squid-2.4.stable1-high_cpu_with_peers.patch
        squid-2.4.stable1-force_valid_blksize.patch
        squid-2.4.stable1-kill_parent_on_child_sigkill.patch
        squid-2.4.stable1-htcp_assertion_fix.patch
        squid-2.4.stable1-diskd_fixed_path.patch

built using the FreeBSD port on FreeBSD 4.3-RC. Thanks!

        -mi
Received on Wed Apr 11 2001 - 07:06:04 MDT