<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Dear Amos,</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
**<span style="color: rgb(0, 0, 0); font-family: Calibri, Helvetica, sans-serif; font-size: 12pt;">i will use each rock dir with one physical disk , i m going to set up it now. + i will change to default rock dir optional values.</span></div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
**Please note that i m switching to rock, since one processor won't hande 800 Mb/s traffic</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
**theoratically would squid with rock cache dir , give me the same gain ratio of ufs cache dir ? [for 100mb/s ufs gave me 70%]</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
**how much do u think of (worker+RAM) is needed for 800 Mb/s traffic?</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
thank u</div>
<div>
<div id="appendonsend"></div>
<div style="font-family:Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
<br>
</div>
<hr tabindex="-1" style="display:inline-block; width:98%">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" color="#000000" style="font-size:11pt"><b>From:</b> squid-users <squid-users-bounces@lists.squid-cache.org> on behalf of Amos Jeffries <squid3@treenet.co.nz><br>
<b>Sent:</b> Thursday, July 2, 2020 1:20 PM<br>
<b>To:</b> squid-users@lists.squid-cache.org <squid-users@lists.squid-cache.org><br>
<b>Subject:</b> Re: [squid-users] rock issue</font>
<div> </div>
</div>
<div class="BodyFragment"><font size="2"><span style="font-size:11pt">
<div class="PlainText">On 2/07/20 8:45 am, patrick mkhael wrote:<br>
> <br>
> ***Please note that you have 20 kids worth mapping (10 workers and 10<br>
> diskers), but you map only the first 10.{since i did not get the point<br>
> of the diskers ,as far as i understood , it should be like (simple<br>
> example)<br>
> >workers 2<br>
>> cpu_affinity_map process_numbers=1,2,3,4 cores=1,2,3,4<br>
>> cache_dir rock /mnt/sdb/1 2048 max-size=10000 max-swap-rate=200 swap-timeout=300<br>
>> cache_dir rock /mnt/sdb/2 2048 max-size=10000 max-swap-rate=200 swap-timeout=300<br>
> <br>
> <br>
> <br>
> ***Why do you have 10 rock caches of various sizes? [ to be honest , i<br>
> saw in many websites that it should be like this from the smallest to<br>
> the bigest with diff size, i tought it should serve from small size pool<br>
> to high ]<br>
> <br>
<br>
In general yes. BUT the size ranges to use should be determined via<br>
traffic analysis. Specifically measure and graph the object sizes being<br>
handled. There will be a wavy / cyclic line resulting. The size<br>
boundaries should be set to the *minimum* point(s) along that line.<br>
<br>
<br>
That said. Since you are comparing the new rock to an old UFS setup. It<br>
would be best to start with the rock being setup as similar to the UFS<br>
as you can - number of cache_dir, ranges of objects stored there etc.<br>
<br>
ie. if these ranges were in the old UFS, then keep them for now. That<br>
can be re-calculated after the HIT ratio drop is identified.<br>
<br>
<br>
> *****How many independent disk spindles (or equivalent) do you have? [ i<br>
> have one raid 5 ssd disks , used by the 10 rock cache dir]<br>
> <br>
<br>
Ouch.<br>
<br>
Ideally you would have either:<br>
<br>
5x SSD disks separately mounted. With one rock cache on each.<br>
<br>
or,<br>
<br>
1x RAID 10 with one rock cache per disk pair/stripe. This requires<br>
ability for controller to map a sub-directory tree exclusively onto one<br>
of the sub-array stripes.<br>
<br>
or,<br>
<br>
2x RAID 1 (drive pair mirroring) with one rock cache on each pair. This<br>
is the simplest way to achieve above when sub-array feature is not<br>
available in RAID 10.<br>
<br>
or,<br>
<br>
1x RAID 10 with a single rock cache<br>
<br>
<br>
<br>
The reasons;<br>
<br>
Squid I/O pattern is mostly writes and erase. Low on reads.<br>
<br>
RAID types in order of best->worst for that pattern are:<br>
none, RAID 1, RAID 10, RAID 5, RAID 0<br>
<<a href="https://wiki.squid-cache.org/SquidFaq/RAID">https://wiki.squid-cache.org/SquidFaq/RAID</a>><br>
<br>
Normal SSD controllers cannot handle the Squid I/O pattern well. Squid<br>
*will* age the disk much faster than manufacturer measured statistics<br>
indicate. (True for even HDD, just less of a problem there).<br>
<br>
This means that the design needs to plan for coping with relatively<br>
frequent disk failures. Loss of the data itself irrelevant. Only the<br>
outage time + reduction in HIT ratio actually matter on failures.<br>
<br>
Look for SSD with high write cycle measurements, and for RAID hot-swap<br>
capability (even if the machine itself cant do that).<br>
<br>
<br>
<br>
> ***How did you select the swap rate limits and timeouts for<br>
> cache_dirs?[I took it also from online forum , can i leave it empty for<br>
> both]<br>
> <br>
<br>
Alex may have better ideas if you can refer us to which tutorials or<br>
documents you found that info in.<br>
<br>
Without specific details on why they were chosen I would go with one<br>
rock cache with default values to start with. Only changing them if<br>
followup analysis indicates some other value is better.<br>
<br>
<br>
Amos<br>
_______________________________________________<br>
squid-users mailing list<br>
squid-users@lists.squid-cache.org<br>
<a href="http://lists.squid-cache.org/listinfo/squid-users">http://lists.squid-cache.org/listinfo/squid-users</a><br>
</div>
</span></font></div>
</div>
</body>
</html>