Re: IMAPD-SSL - too many open files
keithw--- via courier-users <[email protected]> Tue, 27 May 2025 18:48:51 -0500
| Newsgroups | gmane.mail.imap.courier.general |
|---|---|
| Message-ID | <[email protected]> |
Nothing with an insane number of files:
1 /proc/1230/fd
3 /proc/1001696/fd
3 /proc/1001697/fd
3 /proc/1001698/fd
3 /proc/1001699/fd
3 /proc/1001700/fd
3 /proc/798/fd
3 /proc/self/fd
3 /proc/thread-self/fd
4 /proc/1000964/fd
4 /proc/1001014/fd
4 /proc/1001015/fd
4 /proc/1001673/fd
4 /proc/1184/fd
4 /proc/830256/fd
4 /proc/995145/fd
4 /proc/995147/fd
4 /proc/995183/fd
4 /proc/995184/fd
4 /proc/995186/fd
4 /proc/995189/fd
4 /proc/995197/fd
4 /proc/995208/fd
4 /proc/995217/fd
5 /proc/1001607/fd
5 /proc/1001608/fd
5 /proc/1001613/fd
5 /proc/1001615/fd
5 /proc/851/fd
5 /proc/869/fd
5 /proc/995158/fd
5 /proc/995168/fd
5 /proc/995169/fd
5 /proc/995179/fd
5 /proc/995227/fd
5 /proc/995228/fd
5 /proc/995238/fd
5 /proc/995239/fd
5 /proc/995250/fd
5 /proc/995262/fd
6 /proc/1000128/fd
6 /proc/1000366/fd
6 /proc/1000414/fd
6 /proc/1000429/fd
6 /proc/1000508/fd
6 /proc/1000520/fd
6 /proc/1000558/fd
6 /proc/1000634/fd
6 /proc/1000708/fd
6 /proc/1000710/fd
6 /proc/1000711/fd
6 /proc/1000722/fd
6 /proc/1000729/fd
6 /proc/1000738/fd
6 /proc/1000841/fd
6 /proc/1001052/fd
6 /proc/1001108/fd
6 /proc/1001119/fd
6 /proc/1001170/fd
6 /proc/1001195/fd
6 /proc/1001256/fd
6 /proc/1001298/fd
6 /proc/1001324/fd
6 /proc/1001343/fd
6 /proc/1001440/fd
6 /proc/1001456/fd
6 /proc/1001524/fd
6 /proc/1001527/fd
6 /proc/1001561/fd
6 /proc/1001575/fd
6 /proc/1001611/fd
6 /proc/1001660/fd
6 /proc/1001663/fd
6 /proc/1001672/fd
6 /proc/1185/fd
6 /proc/849/fd
6 /proc/870/fd
6 /proc/872/fd
6 /proc/873/fd
6 /proc/874/fd
6 /proc/875/fd
6 /proc/876/fd
6 /proc/877/fd
6 /proc/878/fd
6 /proc/879/fd
6 /proc/880/fd
6 /proc/881/fd
6 /proc/882/fd
6 /proc/883/fd
6 /proc/884/fd
6 /proc/885/fd
6 /proc/886/fd
6 /proc/887/fd
6 /proc/888/fd
6 /proc/889/fd
6 /proc/890/fd
6 /proc/891/fd
6 /proc/892/fd
6 /proc/893/fd
6 /proc/894/fd
6 /proc/895/fd
6 /proc/896/fd
6 /proc/995205/fd
6 /proc/995213/fd
6 /proc/995251/fd
6 /proc/995263/fd
6 /proc/995329/fd
6 /proc/995334/fd
6 /proc/995335/fd
6 /proc/995338/fd
6 /proc/995340/fd
6 /proc/995342/fd
6 /proc/995344/fd
6 /proc/995345/fd
6 /proc/995354/fd
6 /proc/995360/fd
6 /proc/995362/fd
6 /proc/995363/fd
6 /proc/995368/fd
6 /proc/995369/fd
6 /proc/995376/fd
6 /proc/995419/fd
6 /proc/995427/fd
6 /proc/995430/fd
6 /proc/995435/fd
6 /proc/995438/fd
6 /proc/995443/fd
6 /proc/995451/fd
6 /proc/995452/fd
6 /proc/995455/fd
6 /proc/995459/fd
6 /proc/995463/fd
6 /proc/995464/fd
6 /proc/995465/fd
6 /proc/995466/fd
6 /proc/995467/fd
6 /proc/995468/fd
6 /proc/995469/fd
6 /proc/995470/fd
6 /proc/995471/fd
6 /proc/995472/fd
6 /proc/995473/fd
6 /proc/995475/fd
6 /proc/995476/fd
6 /proc/995477/fd
6 /proc/995478/fd
6 /proc/995480/fd
6 /proc/995491/fd
6 /proc/995498/fd
6 /proc/995571/fd
6 /proc/995581/fd
6 /proc/995630/fd
6 /proc/995658/fd
6 /proc/995691/fd
6 /proc/995715/fd
6 /proc/995721/fd
6 /proc/995746/fd
6 /proc/995750/fd
6 /proc/995913/fd
6 /proc/995940/fd
6 /proc/995948/fd
6 /proc/995986/fd
6 /proc/995991/fd
6 /proc/996021/fd
6 /proc/996153/fd
6 /proc/996186/fd
6 /proc/996214/fd
6 /proc/996216/fd
6 /proc/996255/fd
6 /proc/996282/fd
6 /proc/996320/fd
6 /proc/996329/fd
6 /proc/996359/fd
6 /proc/996409/fd
6 /proc/996486/fd
6 /proc/996535/fd
6 /proc/996544/fd
6 /proc/996590/fd
6 /proc/996670/fd
6 /proc/996681/fd
6 /proc/996690/fd
6 /proc/996759/fd
6 /proc/996795/fd
6 /proc/996900/fd
6 /proc/996901/fd
6 /proc/996923/fd
6 /proc/996981/fd
6 /proc/997052/fd
6 /proc/997114/fd
6 /proc/997141/fd
6 /proc/997209/fd
6 /proc/997215/fd
6 /proc/997285/fd
6 /proc/997405/fd
6 /proc/997500/fd
6 /proc/997514/fd
6 /proc/997523/fd
6 /proc/997533/fd
6 /proc/997595/fd
6 /proc/997679/fd
6 /proc/997712/fd
6 /proc/997772/fd
6 /proc/997858/fd
6 /proc/997863/fd
6 /proc/997869/fd
6 /proc/997879/fd
6 /proc/997885/fd
6 /proc/998013/fd
6 /proc/998143/fd
6 /proc/998165/fd
6 /proc/998231/fd
6 /proc/998235/fd
6 /proc/998296/fd
6 /proc/998322/fd
6 /proc/998323/fd
6 /proc/998324/fd
6 /proc/998325/fd
6 /proc/998327/fd
6 /proc/998328/fd
6 /proc/998346/fd
6 /proc/998352/fd
6 /proc/998363/fd
6 /proc/998422/fd
6 /proc/998475/fd
6 /proc/998479/fd
6 /proc/998528/fd
6 /proc/998633/fd
6 /proc/998692/fd
6 /proc/998709/fd
6 /proc/998726/fd
6 /proc/998731/fd
6 /proc/998788/fd
6 /proc/998836/fd
6 /proc/998863/fd
6 /proc/998899/fd
6 /proc/998928/fd
6 /proc/998960/fd
6 /proc/998961/fd
6 /proc/999034/fd
6 /proc/999182/fd
6 /proc/999193/fd
6 /proc/999255/fd
6 /proc/999266/fd
6 /proc/999285/fd
6 /proc/999292/fd
6 /proc/999311/fd
6 /proc/999342/fd
6 /proc/999401/fd
6 /proc/999423/fd
6 /proc/999425/fd
6 /proc/999426/fd
6 /proc/999430/fd
6 /proc/999431/fd
6 /proc/999451/fd
6 /proc/999454/fd
6 /proc/999471/fd
6 /proc/999474/fd
6 /proc/999574/fd
6 /proc/999577/fd
6 /proc/999633/fd
6 /proc/999793/fd
6 /proc/999888/fd
6 /proc/999938/fd
6 /proc/999948/fd
6 /proc/999955/fd
6 /proc/999956/fd
6 /proc/999968/fd
6 /proc/999978/fd
7 /proc/848/fd
7 /proc/865/fd
7 /proc/995162/fd
7 /proc/995164/fd
7 /proc/995166/fd
7 /proc/995171/fd
7 /proc/995174/fd
7 /proc/995194/fd
7 /proc/996545/fd
7 /proc/997115/fd
7 /proc/997886/fd
7 /proc/998727/fd
7 /proc/999267/fd
7 /proc/999286/fd
7 /proc/999294/fd
8 /proc/1000129/fd
8 /proc/1000367/fd
8 /proc/1000431/fd
8 /proc/1000509/fd
8 /proc/1000523/fd
8 /proc/1000559/fd
8 /proc/1000635/fd
8 /proc/1000709/fd
8 /proc/1000723/fd
8 /proc/1000730/fd
8 /proc/1000842/fd
8 /proc/1001053/fd
8 /proc/1001122/fd
8 /proc/1001196/fd
8 /proc/1001257/fd
8 /proc/1001299/fd
8 /proc/1001325/fd
8 /proc/1001344/fd
8 /proc/1001441/fd
8 /proc/1001457/fd
8 /proc/1001562/fd
8 /proc/1001576/fd
8 /proc/1001661/fd
8 /proc/1001664/fd
8 /proc/995188/fd
8 /proc/995331/fd
8 /proc/995337/fd
8 /proc/995341/fd
8 /proc/995347/fd
8 /proc/995348/fd
8 /proc/995349/fd
8 /proc/995352/fd
8 /proc/995356/fd
8 /proc/995364/fd
8 /proc/995366/fd
8 /proc/995371/fd
8 /proc/995372/fd
8 /proc/995378/fd
8 /proc/995421/fd
8 /proc/995428/fd
8 /proc/995432/fd
8 /proc/995436/fd
8 /proc/995440/fd
8 /proc/995444/fd
8 /proc/995453/fd
8 /proc/995457/fd
8 /proc/995458/fd
8 /proc/995460/fd
8 /proc/995479/fd
8 /proc/995481/fd
8 /proc/995482/fd
8 /proc/995483/fd
8 /proc/995484/fd
8 /proc/995485/fd
8 /proc/995486/fd
8 /proc/995487/fd
8 /proc/995488/fd
8 /proc/995489/fd
8 /proc/995492/fd
8 /proc/995494/fd
8 /proc/995495/fd
8 /proc/995496/fd
8 /proc/995497/fd
8 /proc/995499/fd
8 /proc/995500/fd
8 /proc/995503/fd
8 /proc/995582/fd
8 /proc/995632/fd
8 /proc/995722/fd
8 /proc/995914/fd
8 /proc/996022/fd
8 /proc/996154/fd
8 /proc/996215/fd
8 /proc/996257/fd
8 /proc/996322/fd
8 /proc/996331/fd
8 /proc/996410/fd
8 /proc/996487/fd
8 /proc/996536/fd
8 /proc/996591/fd
8 /proc/996671/fd
8 /proc/996682/fd
8 /proc/996691/fd
8 /proc/996760/fd
8 /proc/996796/fd
8 /proc/996902/fd
8 /proc/996903/fd
8 /proc/996924/fd
8 /proc/996982/fd
8 /proc/997054/fd
8 /proc/997142/fd
8 /proc/997211/fd
8 /proc/997217/fd
8 /proc/997287/fd
8 /proc/997406/fd
8 /proc/997524/fd
8 /proc/997534/fd
8 /proc/997597/fd
8 /proc/997680/fd
8 /proc/997713/fd
8 /proc/997773/fd
8 /proc/997859/fd
8 /proc/997864/fd
8 /proc/997870/fd
8 /proc/997880/fd
8 /proc/998014/fd
8 /proc/998144/fd
8 /proc/998166/fd
8 /proc/998232/fd
8 /proc/998236/fd
8 /proc/998297/fd
8 /proc/998326/fd
8 /proc/998330/fd
8 /proc/998347/fd
8 /proc/998353/fd
8 /proc/998423/fd
8 /proc/998529/fd
8 /proc/998634/fd
8 /proc/998693/fd
8 /proc/998710/fd
8 /proc/998732/fd
8 /proc/998789/fd
8 /proc/998837/fd
8 /proc/998865/fd
8 /proc/998902/fd
8 /proc/998929/fd
8 /proc/998963/fd
8 /proc/998964/fd
8 /proc/999035/fd
8 /proc/999183/fd
8 /proc/999194/fd
8 /proc/999260/fd
8 /proc/999313/fd
8 /proc/999343/fd
8 /proc/999403/fd
8 /proc/999424/fd
8 /proc/999434/fd
8 /proc/999435/fd
8 /proc/999452/fd
8 /proc/999634/fd
8 /proc/999794/fd
8 /proc/999889/fd
8 /proc/999939/fd
8 /proc/999950/fd
8 /proc/999957/fd
8 /proc/999958/fd
8 /proc/999969/fd
8 /proc/999980/fd
9 /proc/1000415/fd
9 /proc/1000741/fd
9 /proc/1489/fd
9 /proc/908/fd
9 /proc/995336/fd
9 /proc/995365/fd
9 /proc/995572/fd
9 /proc/995659/fd
9 /proc/995693/fd
9 /proc/995716/fd
9 /proc/995941/fd
9 /proc/995949/fd
9 /proc/995987/fd
9 /proc/995992/fd
9 /proc/996187/fd
9 /proc/996217/fd
9 /proc/996283/fd
9 /proc/996360/fd
9 /proc/997501/fd
9 /proc/997515/fd
9 /proc/998364/fd
9 /proc/998476/fd
9 /proc/998480/fd
9 /proc/999459/fd
9 /proc/999575/fd
9 /proc/999579/fd
10 /proc/1000959/fd
10 /proc/794/fd
10 /proc/796/fd
10 /proc/830241/fd
10 /proc/979095/fd
11 /proc/1001011/fd
11 /proc/669/fd
11 /proc/905/fd
12 /proc/1000963/fd
12 /proc/1001013/fd
12 /proc/909/fd
12 /proc/995159/fd
14 /proc/680/fd
15 /proc/817/fd
17 /proc/995181/fd
23 /proc/844/fd
26 /proc/843/fd
28 /proc/1012/fd
28 /proc/830238/fd
28 /proc/979093/fd
29 /proc/854/fd
33 /proc/627/fd
62 /proc/1013/fd
160 /proc/1/fd
On 2025-05-27 20:18, Sam Varshavchik ([email protected]) wrote:
> keithw--- via courier-users writes:
>
>> After about an hour or two, I start getting this in the logs and then
>> IMAP stops working for everyone.
>>
>> May 20 01:02:45 smtp imapd-ssl[9428]: .: Too many open files
>> May 20 01:02:45 smtp imapd-ssl[9428]: This may be a problem with FAM
>> or Gamin
>> May 20 01:02:45 smtp imapd-ssl[9428]: Failed to create cache file:
>> maildirwatch ([email protected])
>> May 20 01:02:45 smtp imapd-ssl[9428]: Error: Too many open files
>
> Irrespective of anything else, one part of the above message is
> obsolete, and needs to be adjusted.
>
>> A simple restart of courier and the problem is gone.
>>
>> Currently I have a crontab restarting courier hourly and generally it
>> gets us by, but sometimes it doesn't.
>>
>> Here are some relevant info on open file ulimits.
>>
>> root@smtp:/etc/courier# cat /proc/sys/fs/file-max
>> 9223372036854775807
>>
>> root@smtp:/etc/courier# cat /proc/sys/fs/file-nr
>> 6752 0 9223372036854775807
>
> What about /proc/sys/fs/inode-nr and /proc/sys/fs/nr-open ?
>
> https://docs.kernel.org/admin-guide/sysctl/fs.html
>
>> I have a feeling the message is wrong and this is some sort of leak.
>
> Well, one part of the verbiage is wrong, but that is not relevant here.
> But this sounds more like some resource is being exhausted.
>
> On the assumption that this is a file descriptor leak, see if anything
> stands out from:
>
> ls -d /proc/*/fd/* | cut -d/ -f1-4 | sort | uniq -c | sort -n
>
> This counts the number of files opened by each process in the system,
> and sorts the list by number of files.
>
> _______________________________________________
> courier-users mailing list
> [email protected]
> Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
_______________________________________________
courier-users mailing list
[email protected]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users