Re: [squid-users] trouble w/ delay_pools info in cachemgr

Henrik wrote:
> "Current" means how full of bandwidth credit the delay pool is at the
> moment.

So that means if Aggregate Current matches the value for Restore (or Max?)
that means the delay_pool is idle, right? Because no deduction for
user-traffic is occurring? Are the individual IP:CurrPoolVal
realtime/right-now or also include inactives? That is, if a full number
means idle then what are the individual IP:CurrPoolVal entries?
Previous/last used connections or?

Adam asked:
>> Also what does the 1st (lefthand) number in a pair under
>> "Individual" like this "Current: 31:51200 128:45551 38:51200 etc."

and Henrik replied:
> The first is the last component of the IP address. The second the
> current pool value of the individual pool for this IP.

Oops that's pretty obvious, sorry I pick that up. It is possible to hack the
code to show the last *two* octets instead of just the last? We have offices
all over the state and so use a netmask of 255.255.252.0 to divide our IPs.
That means that troubleshooting can give me many of the same 1:### pairs in
the delay pools since the lower numbers are most frequently used at each
site. Just today, troubleshooting the below I got ###.###.6.229 and
###.###.10.229 so if I could get it to report either the full IP (ungainly
but better than nothing) or just the 6.229:### part that would help.

> The pool is not changed if the conditions on which pool the request
> should belong to changes during the duration of the request, but any
> new requests received after the change in conditions will get
> assigned the new pool.

This is what causes me concern/confusion. I have one pool that is based on
pulling .pdf and .doc files more quickly off our offsite/public webserver.
As such it should rarely get used. However I see quite a few users in that
pool. So I checked the logs (for example for 192.168.6.229) and in fact that
user did pull down a few PDF's a few hours ago but since then has only been
surfing craigslist.org. The 5 pdf files they got were all less than a MB.
Surely each download has completed by now (besides I found no instance of
this user's IP in the filedescriptor log which I would have had there been
an open download). And yet each time I go into the delay_pool's menu, that
user's IP shows up. And that pool is *only* for pdf/docs from mysite.

So does the list show active/current delay_pool users or just all those
since X hours?

Here are the delay_pool definitions for that pool:
acl notdelayed1 urlpath_regex -i \.pdf$ \.doc$
acl mysite dstdomain www.mysite.com
http_access allow mysite notdelayed1
delay_pools 4
delay_class 1 2
delay_access 1 allow mysite notdelayed1
delay_access 1 deny all
delay_parameters 1 206000/206000 28000/51200

Similarly, the off-hours pool has a couple of hundred Individual/Current
entries even though that pool is off-hours. Am I doing something wrong or
reading the report incorrectly? Here are the configs for that delay pool
(some dupes with above):
acl delayed time MTWHF 8:01-17:30 <== included as this is the only
other time based acl/delay_pool
acl notdelayed3 time SA 8:01-17:30
acl notdelayed3 time SMTWHFA 17:31-23:59
acl notdelayed3 time SMTWHFA 0:00-8:00
http_access allow localnets notdelayed3
delay_pools 4
delay_class 3 2
delay_access 3 allow notdelayed3
delay_access 3 deny all
delay_access 3 allow notdelayed3
delay_access 3 deny all
delay_parameters 3 206000/206000 32000/51200

thanks again for any answers, advice, pointers, or suggestions.

Adam

P.S.
Here is a typical delay_pool printout (I snipped pools 1,3,4 as the
Indiv/Current lines are quite long).
Delay pools configured: 4
Pool: 1
Class: 2
Aggregate:
Max: 206000
Restore: 206000
Current: 206000
Individual:
Max: 51200
Rate: 28000
Current: 220:51200 4:51200 103:51200 148:51200 191:51200 229:51200 33:51200
Pool: 2
Class: 2
Aggregate:
Max: 206000
Restore: 206000
Current: 206000
Individual:
Max: 51200
Rate: 30000
Current: 1:51200 14:51200 2:51200 91:51200 196:51200
Pool: 3
Class: 2
Aggregate:
Max: 206000
Restore: 206000
Current: 206000
Individual:
Max: 51200
Rate: 32000
Current: 12:51200 132:51200 134:51200 20:51200 101:51200 83:51200
Pool: 4
Class: 2
Aggregate:
Max: 206000
Restore: 206000
Current: 102416
Individual:
Max: 51200
Rate: 16000
Current: 230:51200 116:51200 187:51200 127:51200 117:51200 121:51200
Memory Used: 5440 bytes
Received on Wed May 14 2003 - 15:46:25 MDT