The commit was the latest in slab_rebal_next at the time: https://github.com/dormando/memcached/commit/bdd688b4f20120ad844c8a4803e08c6e03cb061a
addr2line gave me this output: $ addr2line -e memcached 0x40e007 /mnt/builds/slave/workspace/TL-SYS-memcached-slab_rebal_next/build/memcached-1.4.24-slab-rebal-next/slabs.c:264 As well, this was running with production *writes*, but not reads. Even if we had reads on with the few servers crashing, we're ok architecturally. That's why I can get it out there without worrying too much. For now, I'm going to turn it off. I had a metrics issue anyway that needs to get fixed. Tomorrow I'm planning to test again with more metrics, but I can get any new code in pretty quick. On Thursday, October 1, 2015 at 1:01:36 AM UTC-7, Dormando wrote: > > How many servers were you running it on? I hope it wasn't more than a > handful. I'd recommend starting with one :P > > can you do an addr2line? what were your startup args, and what was the > commit sha1 for the branch you pulled? > > sorry about that :/ > > On Thu, 1 Oct 2015, Scott Mansfield wrote: > > > A few different servers (5 / 205) experienced a segfault all within an > hour or so. Unfortunately at this point I'm a bit out of my depth. I have > the dmesg output, which is identical for all 5 boxes: > > > > [46545.316351] memcached[2789]: segfault at 0 ip 000000000040e007 sp > 00007f362ceedeb0 error 4 in memcached[400000+1d000] > > > > > > I can possibly supply the binary file if needed, though we didn't do > anything besides the standard setup and compile. > > > > > > > > On Tuesday, September 29, 2015 at 10:27:59 PM UTC-7, Dormando wrote: > > If you look at the new branch there's a commit explaining the new > stats. > > > > You can watch slab_reassing_evictions vs slab_reassign_saves. you > can also > > test automove=1 vs automove=2 (please also turn on the > lru_maintainer and > > lru_crawler). > > > > The initial branch you were running didn't add any new stats. It > just > > restored an old feature. > > > > On Tue, 29 Sep 2015, Scott Mansfield wrote: > > > > > An unrelated prod problem meant I had to stop after about an > hour. I'm turning it on again tomorrow morning. > > > Are there any new metrics I should be looking at? Anything new > in the stats output? I'm about to take a look at the diffs as well. > > > > > > On Tuesday, September 29, 2015 at 12:37:45 PM UTC-7, Dormando > wrote: > > > excellent. if automove=2 is too aggressive you'll see that > come in in a > > > hit ratio reduction. > > > > > > the new branch works with automove=2 as well, but it will > attempt to > > > rescue valid items in the old slab if possible. I'll still > be working on > > > it for another few hours today though. I'll mail again > when I'm done. > > > > > > On Tue, 29 Sep 2015, Scott Mansfield wrote: > > > > > > > I have the first commit (slab_automove=2) running in > prod right now. Later today will be a full load production test of the > latest code. I'll just let it run for a few days unless I spot any > problems. We have good metrics for latency et. al. from the client side, > though network normally dwarfs memcached time. > > > > > > > > On Tuesday, September 29, 2015 at 3:10:03 AM UTC-7, > Dormando wrote: > > > > That's unfortunate. > > > > > > > > I've done some more work on the branch: > > > > https://github.com/memcached/memcached/pull/112 > > > > > > > > It's not completely likely you would see enough of > an improvement from the > > > > new default mode. However if your item sizes > change gradually, items are > > > > reclaimed during expiration, or get overwritten > (and thus freed in the old > > > > class), it should work just fine. I have another > patch coming which should > > > > help though. > > > > > > > > Open to feedback from any interested party. > > > > > > > > On Fri, 25 Sep 2015, Scott Mansfield wrote: > > > > > > > > > I have it running internally, and it runs fine > under normal load. It's difficult to put it into the line of fire for a > production workload because of social reasons... As well it's a degenerate > case that we normally don't run in to (and actively try to avoid). I'm > going to run some heavier load tests on it today. > > > > > > > > > > On Wednesday, September 9, 2015 at 10:23:32 AM > UTC-7, Scott Mansfield wrote: > > > > > I'm working on getting a test going > internally. I'll let you know how it goes. > > > > > > > > > > > > > > > Scott Mansfield > > > > > On Mon, Sep 7, 2015 at 2:33 PM, dormando wrote: > > > > > Yo, > > > > > > > > > > > https://github.com/dormando/memcached/commits/slab_rebal_next - would you > > > > > mind playing around with the branch here? > You can see the start options in > > > > > the test. > > > > > > > > > > This is a dead simple modification (a > restoration of a feature that was > > > > > arleady there...). The test very > aggressively writes and is able to shunt > > > > > memory around appropriately. > > > > > > > > > > The work I'm exploring right now will > allow savings of items being > > > > > rebalanced from, and increasing the > aggression of page moving without > > > > > being so brain damaged about it. > > > > > > > > > > But while I'm poking around with that, I'd > be interested in knowing if > > > > > this simple branch is an improvement, and > if so how much. > > > > > > > > > > I'll push more code to the branch, but the > changes should be gated behind > > > > > a feature flag. > > > > > > > > > > On Tue, 18 Aug 2015, 'Scott Mansfield' via > memcached wrote: > > > > > > > > > > > > > > > > > No worries man, you're doing us a favor. > Let me know if there's anything you need from us, and I promise I'll be > quicker this time :) > > > > > > > > > > > > On Aug 18, 2015 12:01 AM, "dormando" < > dorm...@rydia.net> wrote: > > > > > > Hey, > > > > > > > > > > > > I'm still really interested in > working on this. I'll be taking a careful > > > > > > look soon I hope. > > > > > > > > > > > > On Mon, 3 Aug 2015, Scott > Mansfield wrote: > > > > > > > > > > > > > I've tweaked the program > slightly, so I'm adding a new version. It prints more stats as it goes and > runs a bit faster. > > > > > > > > > > > > > > On Monday, August 3, 2015 at > 1:20:37 AM UTC-7, Scott Mansfield wrote: > > > > > > > Total brain fart on my > part. Apparently I had memcached 1.4.13 on my path (who knows how...) Using > the actual one that I've built works. Sorry for the confusion... can't > believe I didn't realize that before. I'm testing against the compiled one > now to see how it behaves. > > > > > > > On Monday, August 3, 2015 > at 1:15:06 AM UTC-7, Dormando wrote: > > > > > > > You sure that's > 1.4.24? None of those fail for me :( > > > > > > > > > > > > > > On Mon, 3 Aug 2015, > Scott Mansfield wrote: > > > > > > > > > > > > > > > The command line > I've used that will start is: > > > > > > > > > > > > > > > > memcached -m 64 -o > slab_reassign,slab_automove > > > > > > > > > > > > > > > > > > > > > > > > the ones that fail > are: > > > > > > > > > > > > > > > > > > > > > > > > memcached -m 64 -o > slab_reassign,slab_automove,lru_crawler,lru_maintainer > > > > > > > > > > > > > > > > memcached -o > lru_crawler > > > > > > > > > > > > > > > > > > > > > > > > I'm sure I've > missed something during compile, though I just used ./configure and make. > > > > > > > > > > > > > > > > > > > > > > > > On Monday, August > 3, 2015 at 12:22:33 AM UTC-7, Scott Mansfield wrote: > > > > > > > > I've > attached a pretty simple program to connect, fill a slab with data, and > then fill another slab slowly with data of a different size. I've been > trying to get memcached to run with the lru_crawler and lru_maintainer > flags, but I get ' > > > > > > > > > > > > > > > > Illegal > suboption "(null)"' every time I try to start with either in any > configuration. > > > > > > > > > > > > > > > > > > > > > > > > I haven't > seen it start to move slabs automatically with a freshly installed 1.2.24. > > > > > > > > > > > > > > > > > > > > > > > > On Tuesday, > July 21, 2015 at 4:55:17 PM UTC-7, Scott Mansfield wrote: > > > > > > > > I > realize I've not given you the tests to reproduce the behavior. I should be > able to soon. Sorry about the delay here. > > > > > > > > In the mean time, > I wanted to bring up a possible secondary use of the same logic to move > items on slab rebalancing. I think the system might benefit from using the > same logic to crawl the pages in a slab and compact the data in the > background. In the case where we have memory that is assigned to > > the slab > > > but not > > > > > being used > > > > > > because > > > > > > > of replaced > > > > > > > > or TTL'd out data, > returning the memory to a pool of free memory will allow a slab to grow > with that memory first instead of waiting for an event where memory is > needed at that instant. > > > > > > > > > > > > > > > > It's a change in > approach, from reactive to proactive. What do you think? > > > > > > > > > > > > > > > > On Monday, July > 13, 2015 at 5:54:11 PM UTC-7, Dormando wrote: > > > > > > > > > First, > more detail for you: > > > > > > > > > > > > > > > > > > We are > running 1.4.24 in production and haven't noticed any bugs as of yet. The > new LRUs seem to be working well, though we nearly always run memcached > scaled to hold all data without evictions. Those with evictions are > behaving well. Those without evictions haven't seen crashing or any > > other > > > noticeable > > > > bad > > > > > behavior. > > > > > > > > > > > > > > > > Neat. > > > > > > > > > > > > > > > > > > > > > > > > > > OK, I > think I see an area where I was speculating on functionality. If you have a > key in slab 21 and then the same key is written again at a larger size in > slab 23 I assumed that the space in 21 was not freed on the second write. > With that assumption, the LRU crawler would not free up that > > space. > > > Also just > > > > > by observation > > > > > > in > > > > > > > the > > > > > > > > macro, the > space is not freed > > > > > > > > > fast > enough to be effective, in our use case, to accept the writes that are > happening. Think in the hundreds of millions of "overwrites" in a 6 - 10 > hour period across a cluster. > > > > > > > > > > > > > > > > Internally, > "items" (a key/value pair) are generally immutable. The only > > > > > > > > time when > it's not is for INCR/DECR, and it still becomes immutable if two > > > > > > > > INCR/DECR's > collide. > > > > > > > > > > > > > > > > What this > means, is that the new item is staged in a piece of free memory > > > > > > > > while the > "upload" stage of the SET happens. When memcached has all of the > > > > > > > > data in > memory to replace the item, it does an internal swap under a lock. > > > > > > > > The old item > is removed from the hash table and LRU, and the new item gets > > > > > > > > put in its > place (at the head of the LRU). > > > > > > > > > > > > > > > > Since items > are refcounted, this means that if other users are downloading > > > > > > > > an item > which just got replaced, their memory doesn't get corrupted by the > > > > > > > > item > changing out from underneath them. They can continue to read the old > > > > > > > > item until > they're done. When the refcount reaches zero the old memory is > > > > > > > > reclaimed. > > > > > > > > > > > > > > > > Most of the > time, the item replacement happens then the old memory is > > > > > > > > immediately > removed. > > > > > > > > > > > > > > > > However, > this does mean that you need *one* piece of free memory to > > > > > > > > replace the > old one. Then the old memory gets freed after that set. > > > > > > > > > > > > > > > > So if you > take a memcached instance with 0 free chunks, and do a rolling > > > > > > > > replacement > of all items (within the same slab class as before), the first > > > > > > > > one would > cause an eviction from the tail of the LRU to get a free chunk. > > > > > > > > Every SET > after that would use the chunk freed from the replacement of the > > > > > > > > previous > memory. > > > > > > > > > > > > > > > > > After that > last sentence I realized I also may not have explained well enough the > access pattern. The keys are all overwritten every day, but it takes some > time to write them all (obviously). We see a huge increase in the bytes > metric as if the new data for the old keys was being written > > for the > > > first > > > > time. > > > > > Since the > > > > > > "old" > > > > > > > slab for > > > > > > > > the same key > doesn't > > > > > > > > > > proactively release memory, it starts to fill up the cache and then start > evicting data in the new slab. Once that happens, we see evictions in the > old slab because of the algorithm you mentioned (random picking / freeing > of memory). Typically we don't see any use for "upgrading" an item as > > the new > > > data > > > > > would be entirely > > > > > > > new and > > > > > > > > should > wholesale replace the > > > > > > > > > old data > for that key. More specifically, the operation is always set, with > different data each day. > > > > > > > > > > > > > > > > Right. Most > of your problems will come from two areas. One being that > > > > > > > > writing data > aggressively into the new slab class (unless you set the > > > > > > > > rebalancer > to always-replace mode), the mover will make memory available > > > > > > > > more slowly > than you can insert. So you'll cause extra evictions in the > > > > > > > > new slab > class. > > > > > > > > > > > > > > > > The > secondary problem is from the random evictions in the previous slab > > > > > > > > class as > stuff is chucked on the floor to make memory moveable. > > > > > > > > > > > > > > > > > As for > testing, we'll be able to put it under real production workload. I don't > know what kind of data you mean you need for testing. The data stored in > the caches are highly confidential. I can give you all kinds of metrics, > since we collect most of the ones that are in the stats and some > > from the > > > stats > > > > > slabs output. If > > > > > > > you have > > > > > > > > some > specific ones that > > > > > > > > > need > collecting, I'll double check and make sure we can get those. > Alternatively, it might be most beneficial to see the metrics in person :) > > > > > > > > > > > > > > > > I just need > stats snapshots here and there, and actually putting the thing > > > > > > > > under load. > When I did the LRU work I had to beg for several months > > > > > > > > before > anyone tested it with a production load. This slows things down and > > > > > > > > demotivates > me from working on the project. > > > > > > > > > > > > > > > > > Unfortunately my dayjob keeps me pretty busy so ~internet~ would probably > > > > > > > > be best. > > > > > > > > > > > > > > > > > I can > create a driver program to reproduce the behavior on a smaller scale. It > would write e.g. 10k keys of 10k size, then rewrite the same keys with > different size data. I'll work on that and post it to this thread when I > can reproduce the behavior locally. > > > > > > > > > > > > > > > > Ok. There're > slab rebalance unit tests in the t/ directory which do things > > > > > > > > like this, > and I've used mc-crusher to slam the rebalancer. It's pretty > > > > > > > > easy to run > one config to load up 10k objects, then flip to the other > > > > > > > > using the > same key namespace. > > > > > > > > > > > > > > > > > Thanks, > > > > > > > > > Scott > > > > > > > > > > > > > > > > > > On > Saturday, July 11, 2015 at 12:05:54 PM UTC-7, Dormando wrote: > > > > > > > > > Hey, > > > > > > > > > > > > > > > > > > On > Fri, 10 Jul 2015, Scott Mansfield wrote: > > > > > > > > > > > > > > > > > > > > We've seen issues recently where we run a cluster that typically has the > majority of items overwritten in the same slab every day and a sudden > change in data size evicts a ton of data, affecting downstream systems. To > be clear that is our problem, but I think there's a tweak in > > memcached > > > that might > > > > > be useful and > > > > > > > another > > > > > > > > possible > feature that > > > > > > > > > > would be even > > > > > > > > > > > better. > > > > > > > > > > > The data that is written to this cache is overwritten every day, though the > TTL is 7 days. One slab takes up the majority of the space in the cache. > The application wrote e.g. 10KB (slab 21) every day for each key > consistently. One day, a change occurred where it started writing > > 15KB (slab > > > 23), > > > > > causing a migration > > > > > > > of data > > > > > > > > from one > slab to > > > > > > > > > > another. We had -o > > > > > > > > > > > slab_reassign,slab_automove=1 set on the server, causing large numbers of > evictions on the initial slab. Let's say the cache could hold the data at > 15KB per key, but the old data was not technically TTL'd out in it's old > slab. This means that memory was not being freed by the lru > > crawler > > > thread (I > > > > > think) because > > > > > > its > > > > > > > expiry > > > > > > > > had not come > > > > > > > > > > around. > > > > > > > > > > > > > > > > > > > > > lines 1199 and 1200 in items.c: > > > > > > > > > > if > ((search->exptime != 0 && search->exptime < current_time) || > is_flushed(search)) { > > > > > > > > > > > > > > > > > > > > If > there was a check to see if this data was "orphaned," i.e. that the key, if > accessed, would map to a different slab than the current one, then these > orphans could be reclaimed as free memory. I am working on a patch to do > this, though I have reservations about performing a hash > > on the > > > key on the > > > > > lru crawler > > > > > > > thread (if > > > > > > > > the hash is > not > > > > > > > > > > already available). > > > > > > > > > > I > have very little experience in the memcached codebase so I don't know the > most efficient way to do this. Any help would be appreciated. > > > > > > > > > > > > > > > > > > > There seems to be a misconception about how the slab classes work. A key, > > > > > > > > > if > already existing in a slab, will always map to the slab class it > > > > > > > > > > currently fits into. The slab classes always exist, but the amount of > > > > > > > > > > memory reserved for each of them will shift with the slab_reassign. ie: 10 > > > > > > > > > > pages in slab class 21, then memory pressure on 23 causes it to move over. > > > > > > > > > > > > > > > > > > So > if you examine a key that still exists in slab class 21, it has no > > > > > > > > > > reason to move up or down the slab classes. > > > > > > > > > > > > > > > > > > > > Alternatively, and possibly more beneficial is compaction of data in a slab > using the same set of criteria as lru crawling. Understandably, compaction > is a very difficult problem to solve since moving the data would be a pain > in the ass. I saw a couple of discussions about this in > > the > > > mailing list, > > > > > though I didn't > > > > > > > see any > > > > > > > > firm > thoughts about > > > > > > > > > it. > I think it > > > > > > > > > > > can probably be done in O(1) like the lru crawler by limiting the number of > items it touches each time. Writing and reading are doable in O(1) so > moving should be as well. Has anyone given more thought on compaction? > > > > > > > > > > > > > > > > > > I'd > be interested in hacking this up for you folks if you can provide me > > > > > > > > > > testing and some data to work with. With all of the LRU work I did in > > > > > > > > > > 1.4.24, the next things I wanted to do is a big improvement on the slab > > > > > > > > > > reassignment code. > > > > > > > > > > > > > > > > > > > Currently it picks essentially a random slab page, empties it, and moves > > > > > > > > > the > slab page into the class under pressure. > > > > > > > > > > > > > > > > > > One > thing we can do is first examine for free memory in the existing slab, > > > > > > > > > IE: > > > > > > > > > > > > > > > > > > - > Take a page from slab 21 > > > > > > > > > - > Scan the page for valid items which need to be moved > > > > > > > > > - > Pull free memory from slab 21, migrate the item (moderately complicated) > > > > > > > > > - > When the page is empty, move it (or give up if you run out of free > > > > > > > > > > chunks). > > > > > > > > > > > > > > > > > > The > next step is to pull from the LRU on slab 21: > > > > > > > > > > > > > > > > > > - > Take page from slab 21 > > > > > > > > > - > Scan page for valid items > > > > > > > > > - > Pull free memory from slab 21, migrate the item > > > > > > > > > - > If no memory free, evict tail of slab 21. use that chunk. > > > > > > > > > - > When the page is empty, move it. > > > > > > > > > > > > > > > > > > > Then, when you hit this condition your least-recently-used data gets > > > > > > > > > > culled as new data migrates your page class. This should match a natural > > > > > > > > > > occurrance if you would already be evicting valid (but old) items to make > > > > > > > > > room > for new items. > > > > > > > > > > > > > > > > > > A > bonus to using the free memory trick, is that I can use the amount of > > > > > > > > > free > space in a slab class as a heuristic to more quickly move slab pages > > > > > > > > > > around. > > > > > > > > > > > > > > > > > > If > it's still necessary from there, we can explore "upgrading" items to a > > > > > > > > > new > slab class, but that is much much more complicated since the item has > > > > > > > > > to > shift LRU's. Do you put it at the head, the tail, the middle, etc? It > > > > > > > > > > might be impossible to make a good generic decision there. > > > > > > > > > > > > > > > > > > What > version are you currently on? If 1.4.24, have you seen any > > > > > > > > > > instability? I'm currently torn between fighting a few bugs and start on > > > > > > > > > > improving the slab rebalancer. > > > > > > > > > > > > > > > > > > > -Dormando > > > > > > > > > > > > > > > > > > > > > > > > > > > On > Saturday, July 11, 2015 at 12:05:54 PM UTC-7, Dormando wrote: > > > > > > > > > Hey, > > > > > > > > > > > > > > > > > > On > Fri, 10 Jul 2015, Scott Mansfield wrote: > > > > > > > > > > > > > > > > > > > > We've seen issues recently where we run a cluster that typically has the > majority of items overwritten in the same slab every day and a sudden > change in data size evicts a ton of data, affecting downstream systems. To > be clear that is our problem, but I think there's a tweak in > > memcached > > > that might > > > > > be useful and > > > > > > > another > > > > > > > > possible > feature that > > > > > > > > > > would be even > > > > > > > > > > > better. > > > > > > > > > > > The data that is written to this cache is overwritten every day, though the > TTL is 7 days. One slab takes up the majority of the space in the cache. > The application wrote e.g. 10KB (slab 21) every day for each key > consistently. One day, a change occurred where it started writing > > 15KB (slab > > > 23), > > > > > causing a migration > > > > > > > of data > > > > > > > > from one > slab to > > > > > > > > > > another. We had -o > > > > > > > > > > > slab_reassign,slab_automove=1 set on the server, causing large numbers of > evictions on the initial slab. Let's say the cache could hold the data at > 15KB per key, but the old data was not technically TTL'd out in it's old > slab. This means that memory was not being freed by the lru > > crawler > > > thread (I > > > > > think) because > > > > > > its > > > > > > > expiry > > > > > > > > had not come > > > > > > > > > > around. > > > > > > > > > > > > > > > > > > > > > lines 1199 and 1200 in items.c: > > > > > > > > > > if > ((search->exptime != 0 && search->exptime < current_time) || > is_flushed(search)) { > > > > > > > > > > > > > > > > > > > > If > there was a check to see if this data was "orphaned," i.e. that the key, if > accessed, would map to a different slab than the current one, then these > orphans could be reclaimed as free memory. I am working on a patch to do > this, though I have reservations about performing a hash > > on the > > > key on the > > > > > lru crawler > > > > > > > thread (if > > > > > > > > the hash is > not > > > > > > > > > > already available). > > > > > > > > > > I > have very little experience in the memcached codebase so I don't know the > most efficient way to do this. Any help would be appreciated. > > > > > > > > > > > > > > > > > > > There seems to be a misconception about how the slab classes work. A key, > > > > > > > > > if > already existing in a slab, will always map to the slab class it > > > > > > > > > > currently fits into. The slab classes always exist, but the amount of > > > > > > > > > > memory reserved for each of them will shift with the slab_reassign. ie: 10 > > > > > > > > > > pages in slab class 21, then memory pressure on 23 causes it to move over. > > > > > > > > > > > > > > > > > > So > if you examine a key that still exists in slab class 21, it has no > > > > > > > > > > reason to move up or down the slab classes. > > > > > > > > > > > > > > > > > > > > Alternatively, and possibly more beneficial is compaction of data in a slab > using the same set of criteria as lru crawling. Understandably, compaction > is a very difficult problem to solve since moving the data would be a pain > in the ass. I saw a couple of discussions about this in > > the > > > mailing list, > > > > > though I didn't > > > > > > > see any > > > > > > > > firm > thoughts about > > > > > > > > > it. > I think it > > > > > > > > > > > can probably be done in O(1) like the lru crawler by limiting the number of > items it touches each time. Writing and reading are doable in O(1) so > moving should be as well. Has anyone given more thought on compaction? > > > > > > > > > > > > > > > > > > I'd > be interested in hacking this up for you folks if you can provide me > > > > > > > > > > testing and some data to work with. With all of the LRU work I did in > > > > > > > > > > 1.4.24, the next things I wanted to do is a big improvement on the slab > > > > > > > > > > reassignment code. > > > > > > > > > > > > > > > > > > > Currently it picks essentially a random slab page, empties it, and moves > > > > > > > > > the > slab page into the class under pressure. > > > > > > > > > > > > > > > > > > One > thing we can do is first examine for free memory in the existing slab, > > > > > > > > > IE: > > > > > > > > > > > > > > > > > > - > Take a page from slab 21 > > > > > > > > > - > Scan the page for valid items which need to be moved > > > > > ... -- --- You received this message because you are subscribed to the Google Groups "memcached" group. To unsubscribe from this group and stop receiving emails from it, send an email to memcached+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
