Help Fine Tuning Advanced Cache Settings
-
Hi,
I'm on the advanced cache page and I'm seeing Post Cache, Object Cache, Group Cache and Local Cache are all maxed out.
I'm guessing this affects performance somehow, right?
I am able to tweak Post Cache Size. I'm trying to double it right now. I'm assuming this would improve performance as we'll be getting more Hits and Misses.
I am not able to tweak the size for the other 3. Is that intentional?
Thanks!
Best regards,
Pat -
The caches should be maxed out after a while, that's intended. When the cache is full the least recently used entries are dropped when a new item is added to the cache.
The default values are fine for most installs and only the first one is configurable from the ACP. That page should show you the hit ratio for each cache. The higher the hit ratio the better. Here are the stats for this forum.
The local cache might not max out depending on how many categories you have. That cache is used for certain database queries like loading the category list or the children of a category. Since these rarely change and the hit ratio there is almost 1 which means we only query the database if you change categories.
-
Can you post your stats? How many categories/groups do you have?
The group cache is used to store the membership of users in groups(and privileges). For example checking if a user is an administrator is done by checking if they are member of the
administrators
group. The results are cached since group membership rarely changes.
If the hit ratio is low you might have lots of active registered users and/or only a single nodebb instance. Each nodebb process has it's own cache and if users are distributed to these nodebb instances evenly the group cache will have a higher hit ratio. -
The toggle lets you enable/disable the cache, the delete button just clears it's contents.
You can download the contents of the cache as well.
When you clear the group cache how long does it take to make that many misses?
How many users are on your nodebb? You can check /admin/development/info for more detailed stats.
My guess is there is some code that is making a lot of group membership checks maybe a plugin. Or you just have a lot of active users.
-
ip-10-111-250-67:4567 / 7985 v10.17.0 82 / 4 / 93 v1.16.x@d7044a 1244.30 / 39.95 292.63 mb / 10635.03 mb 1.05, 1.64, 1.87 57m ip-10-111-250-67:4568 / 7986 v10.17.0 89 / 7 / 106 v1.16.x@d7044a 1176.16 / 38.71 340.14 mb / 10650.05 mb 1.05, 1.64, 1.87 57m ip-10-111-250-67:4569 / 7988 v10.17.0 106 / 12 / 150 v1.16.x@d7044a 1395.36 / 43.34 369.73 mb / 10650.05 mb 1.05, 1.64, 1.87 57m ip-10-111-250-67:4570 / 8005 v10.17.0 97 / 5 / 123 v1.16.x@d7044a 1431.64 / 45.03 312.77 mb / 10650.05 mb 1.05, 1.64, 1.87 57m
[ { "process": { "port": "4567", "pid": 7985, "title": "/usr/bin/node", "version": "v10.17.0", "memoryUsage": { "rss": 331522048, "heapTotal": 270622720, "heapUsed": 191571224, "external": 23467064, "humanReadable": "316.16" }, "uptime": 3516.246, "cpuUsage": { "user": "1258.65", "system": "40.50" }, "uptimeHumanReadable": "58m" }, "os": { "hostname": "ip-10-111-250-67", "type": "Linux", "platform": "linux", "arch": "x64", "release": "5.4.0-1029-aws", "load": "1.70, 1.72, 1.89", "freemem": "10625.66", "totalmem": "16481.63" }, "nodebb": { "isCluster": true, "isPrimary": true, "runJobs": true, "jobsDisabled": false }, "git": { "hash": "d7044a5e8623ae6b95360ff58e8f6dbc0e90629b", "hashShort": "d7044a", "branch": "v1.16.x" }, "stats": { "onlineGuestCount": 4, "onlineRegisteredCount": 82, "socketCount": 93, "users": { "categories": 0, "recent": 1, "unread": 0, "topics": 39, "category": 41 }, "topics": [ { "tid": "26200", "count": 5 }, { "tid": "26202", "count": 4 }, { "tid": "26203", "count": 3 }, { "tid": "26747", "count": 3 }, { "tid": "26359", "count": 2 }, { "tid": "26199", "count": 2 }, { "tid": "26211", "count": 2 }, { "tid": "26901", "count": 2 }, { "tid": "26634", "count": 1 }, { "tid": "25866", "count": 1 } ] }, "id": "ip-10-111-250-67:4567" }, { "process": { "port": "4568", "pid": 7986, "title": "/usr/bin/node", "version": "v10.17.0", "memoryUsage": { "rss": 370384896, "heapTotal": 302641152, "heapUsed": 261548072, "external": 22358009, "humanReadable": "353.23" }, "uptime": 3516.239, "cpuUsage": { "user": "1204.29", "system": "39.61" }, "uptimeHumanReadable": "58m" }, "os": { "hostname": "ip-10-111-250-67", "type": "Linux", "platform": "linux", "arch": "x64", "release": "5.4.0-1029-aws", "load": "1.70, 1.72, 1.89", "freemem": "10625.66", "totalmem": "16481.63" }, "nodebb": { "isCluster": true, "isPrimary": false, "runJobs": false, "jobsDisabled": false }, "git": { "hash": "d7044a5e8623ae6b95360ff58e8f6dbc0e90629b", "hashShort": "d7044a", "branch": "v1.16.x" }, "stats": { "onlineGuestCount": 5, "onlineRegisteredCount": 93, "socketCount": 110, "users": { "categories": 2, "recent": 2, "unread": 0, "topics": 41, "category": 48 }, "topics": [ { "tid": "26747", "count": 7 }, { "tid": "26201", "count": 6 }, { "tid": "26200", "count": 3 }, { "tid": "26202", "count": 2 }, { "tid": "26922", "count": 1 }, { "tid": "26359", "count": 1 }, { "tid": "26794", "count": 1 }, { "tid": "25999", "count": 1 }, { "tid": "26210", "count": 1 }, { "tid": "26136", "count": 1 } ] }, "id": "ip-10-111-250-67:4568" }, { "process": { "port": "4569", "pid": 7988, "title": "/usr/bin/node", "version": "v10.17.0", "memoryUsage": { "rss": 365465600, "heapTotal": 299364352, "heapUsed": 262696736, "external": 23254567, "humanReadable": "348.54" }, "uptime": 3516.235, "cpuUsage": { "user": "1411.07", "system": "43.92" }, "uptimeHumanReadable": "58m" }, "os": { "hostname": "ip-10-111-250-67", "type": "Linux", "platform": "linux", "arch": "x64", "release": "5.4.0-1029-aws", "load": "1.70, 1.72, 1.89", "freemem": "10625.66", "totalmem": "16481.63" }, "nodebb": { "isCluster": true, "isPrimary": false, "runJobs": false, "jobsDisabled": false }, "git": { "hash": "d7044a5e8623ae6b95360ff58e8f6dbc0e90629b", "hashShort": "d7044a", "branch": "v1.16.x" }, "stats": { "onlineGuestCount": 12, "onlineRegisteredCount": 105, "socketCount": 151, "users": { "categories": 2, "recent": 4, "unread": 0, "topics": 69, "category": 48 }, "topics": [ { "tid": "26200", "count": 8 }, { "tid": "26747", "count": 8 }, { "tid": "26201", "count": 3 }, { "tid": "26656", "count": 3 }, { "tid": "26927", "count": 3 }, { "tid": "26533", "count": 2 }, { "tid": "26209", "count": 2 }, { "tid": "26199", "count": 2 }, { "tid": "26727", "count": 2 }, { "tid": "26901", "count": 2 } ] }, "id": "ip-10-111-250-67:4569" }, { "process": { "port": "4570", "pid": 8005, "title": "/usr/bin/node", "version": "v10.17.0", "memoryUsage": { "rss": 336445440, "heapTotal": 272719872, "heapUsed": 218833992, "external": 22805176, "humanReadable": "320.86" }, "uptime": 3516.232, "cpuUsage": { "user": "1454.85", "system": "45.98" }, "uptimeHumanReadable": "58m" }, "os": { "hostname": "ip-10-111-250-67", "type": "Linux", "platform": "linux", "arch": "x64", "release": "5.4.0-1029-aws", "load": "1.70, 1.72, 1.89", "freemem": "10616.37", "totalmem": "16481.63" }, "nodebb": { "isCluster": true, "isPrimary": false, "runJobs": false, "jobsDisabled": false }, "git": { "hash": "d7044a5e8623ae6b95360ff58e8f6dbc0e90629b", "hashShort": "d7044a", "branch": "v1.16.x" }, "stats": { "onlineGuestCount": 5, "onlineRegisteredCount": 101, "socketCount": 128, "users": { "categories": 1, "recent": 2, "unread": 0, "topics": 63, "category": 46 }, "topics": [ { "tid": "26200", "count": 10 }, { "tid": "26747", "count": 6 }, { "tid": "26201", "count": 6 }, { "tid": "26199", "count": 5 }, { "tid": "26359", "count": 4 }, { "tid": "26203", "count": 4 }, { "tid": "26210", "count": 3 }, { "tid": "26728", "count": 2 }, { "tid": "26656", "count": 2 }, { "tid": "26204", "count": 2 } ] }, "id": "ip-10-111-250-67:4570" } ]
-
@baris Can I increase group cache size? just curious.
I'm reviewing what's on the cache dump. Looks like lots of permission stuff. We have multiple moderators per category. Maybe that's adding up to that list. I can probably clean that up and see how it looks after.
-
@baris said in Help Fine Tuning Advanced Cache Settings:
@baris said in Help Fine Tuning Advanced Cache Settings:
/admin/development/info
New one for me. And pretty nice. Is this documented somewhere? Or is it one of those top seekrit dev thangs...
Thx--