Re: [squid-users] RockStore "Fatal Error"

From: Nyamul Hassan <nyamul_at_gmail.com>
Date: Mon, 7 Jul 2014 05:13:11 +0600

All permissions are Squid.Squid:

[root_at_proxy01 scripts]# ll /cachestore/cache[1-4]/rock/
/cachestore/cache1/rock/:
total 16
-rw------- 1 squid squid 21474836480 Jul 7 01:48 rock

/cachestore/cache2/rock/:
total 16
-rw------- 1 squid squid 21474836480 Jul 7 01:48 rock

/cachestore/cache3/rock/:
total 16
-rw------- 1 squid squid 21474836480 Jul 7 01:48 rock

/cachestore/cache4/rock/:
total 16
-rw------- 1 squid squid 21474836480 Jul 7 01:48 rock

Tested with both 3.4.6 & 3.4.5-2 (that I got from you)

Regards
HASSAN

On Mon, Jul 7, 2014 at 4:16 AM, Eliezer Croitoru <eliezer_at_ngtech.co.il> wrote:
> What is the permissions of the /cachestore/cache1/rock/rock ?
> "ls -laz /cachestore/cache1/rock/rock" should show some details?
>
> What version of squid are you using?
> squid -v?
>
> Eliezer
>
>
> On 07/07/2014 01:14 AM, Nyamul Hassan wrote:
>>
>> Hi,
>>
>> We were trying to configure the new RockStore as a replacement to the
>> COSS from Squid 2.7. We are seeing these lines in "cache.log"
>>
>> FATAL: Rock cache_dir at /cachestore/cache1/rock/rock failed to open
>> db file: (11) Resource temporarily unavailable
>>
>> Full cache.log is here:http://pastebin.com/dWD2di48
>>
>> We have 4 RockStore cache_dir lines like the one below:
>>
>> cache_dir rock /cachestore/cache1/rock 20480 max-size=32767
>> swap-timeout=300 max-swap-rate=300
>>
>> Full squid.conf is here:http://pastebin.com/cQEWZTuf
>>
>> We did a full umount, mkfs.ext4, mount & did a "-z" to recreate the
>> cache_dir. That "-z" did not take any time at all, hardly 2-3
>> seconds.
>>
>> If I disable the 1st RockStore config line, then the 2nd gives the
>> same error, so and so forth.
>>
>> Can someone help?
>>
>> Regards
>> HASSAN
>>
>
Received on Sun Jul 06 2014 - 23:13:57 MDT

This archive was generated by hypermail 2.2.0 : Mon Jul 07 2014 - 12:00:04 MDT